feat: add OpenAI-compat /v1/chat/completions endpoint (TODO #22)

embedded a screenshot (hopefully) into the text

.gitignore

@@ -4,3 +4,4 @@
 __pycache__/
 venv/
 readme.md-
+*.bak

README.md

@@ -0,0 +1,453 @@
+![jarvisChat logo](static/jcscreenie.png)
+# ⚡ JarvisChat v1.9.0
+
+**A privacy-first, homelab-native developer knowledge platform.**
+
+> JarvisChat turns a heterogeneous LAN of budget hardware into a distributed local AI inference cluster — accumulating institutional knowledge over time, keeping all data off the cloud, and squeezing real performance out of modest consumer hardware through architecture rather than dollars.
+
+This is not another AI chat wrapper. jC is the UX and knowledge-management layer for a local AI brain — analogous to what Windows was to DOS, or what the web is to the internet. The intelligence lives in the model and the RAG corpus. jC makes it accessible and keeps feeding it.
+
+---
+
+## The Four Pillars
+
+### 1. Privacy
+Everything runs on your LAN. No API keys, no cloud endpoints, no data leaving your network, no subscription, no terms-of-service surprises. Your conversations, your codebase, your decisions — stay yours.
+
+### 2. Knowledge Retention
+Unlike stateless chat tools that forget everything when you close the tab, jC accumulates institutional memory. Every solved problem, every architectural decision, every working command gets absorbed into the RAG corpus via Qdrant. The system gets smarter the longer you use it.
+
+### 3. Budget Hardware Maximization
+You don't need a $10,000 workstation. jC is designed for the developer who has a drawer full of machines and the skills to wire them together. RPC clustering, model splitting across CPU and GPU nodes, dynamic resource negotiation, and smart RAG eviction squeeze real performance out of modest consumer hardware.
+
+### 4. Homelab-Native Architecture
+Built specifically for the heterogeneous homelab: mixed hardware, mixed OS, consumer GPUs, ARM boards, NAS storage — all working together as a coherent AI platform. A designated master node hosts jC, llama-server, and SearXNG. GPU nodes self-register as RPC inference workers. The architecture scales horizontally across whatever you've got.
+
+---
+
+## Target Audience
+
+Solo developers and homelab enthusiasts who are:
+- Budget-constrained but hardware-rich (multiple machines, NAS, spare GPUs)
+- Privacy-conscious (no cloud AI subscriptions)
+- Technically capable (if you can install jC, you can designate the master node)
+- Building something over time and want their AI to remember it
+
+---
+
+## Architecture
+
+```
+┌─────────────────────────────────────────────────────────────┐
+│                        YOUR LAN                             │
+│                                                             │
+│  ┌─────────────────┐         ┌──────────────────────────┐  │
+│  │   jarvis        │◄──RPC───│   ultron                 │  │
+│  │   192.168.50.212│  50052  │   192.168.50.108         │  │
+│  │                 │         │                          │  │
+│  │  jC :8080       │         │  llama-server :8081      │  │
+│  │  SearXNG :8888  │         │  llama-server :8082 (*)  │  │
+│  │  RX 6600 XT 8GB │         │  Qdrant :6333            │  │
+│  │  GPU RPC worker │         │  mxbai-embed :11434      │  │
+│  │  Vulkan backend │         │  AMD Ryzen 7 7840HS      │  │
+│  └─────────────────┘         │  Radeon 780M iGPU        │  │
+│                              └──────────────────────────┘  │
+│                                                             │
+│  ┌─────────────────┐         ┌──────────────────────────┐  │
+│  │   pivault       │         │   corsair                │  │
+│  │   192.168.50.158│         │   192.168.50.132         │  │
+│  │                 │         │                          │  │
+│  │  10.83TB RAID5  │         │  RTX 5070 Ti 16GB        │  │
+│  │  RPi 5 8GB      │         │  Ryzen 7 7800X3D         │  │
+│  │  NAS / Kopia    │         │  Gaming / Streaming      │  │
+│  └─────────────────┘         └──────────────────────────┘  │
+│                                                             │
+│  (*) Planned: Qwen2.5-Coder-14B on :8082                   │
+└─────────────────────────────────────────────────────────────┘
+```
+
+**Data flow:**
+```
+Browser / IDE (Continue.dev)
+    → jC :8080 (FastAPI — auth, RAG, memory, conversation history)
+        → Qdrant :6333 (vector search, mxbai-embed-large for embeddings)
+        → llama-server :8081 (inference)
+            → jarvis RPC :50052 (GPU layer offload — RX 6600 XT)
+```
+
+---
+
+## The AMD + NVIDIA Cross-Cluster Reality
+
+This cluster intentionally mixes GPU architectures — **AMD RX 6600 XT on jarvis** and **NVIDIA RTX 5070 Ti on corsair**. This is deliberate and it works.
+
+The RPC layer in llama.cpp is GPU-vendor-agnostic. jarvis runs llama-rpc with a **Vulkan backend** (not ROCm, not CUDA) which provides hardware-neutral GPU acceleration. ultron's llama-server connects to it over TCP and offloads tensor layers without caring what GPU is on the other end.
+
+This means any machine on your LAN with any GPU (AMD, NVIDIA, Intel Arc) can participate as an RPC worker — as long as it can run llama-rpc with Vulkan support.
+
+---
+
+## Cluster Performance Tuning
+
+### The Layer Offloading Trick
+
+The key to squeezing performance out of a CPU+GPU split cluster is `--n-gpu-layers`. This controls how many transformer layers get offloaded to the RPC GPU backend versus staying on the CPU.
+
+**Starting point (before tuning):** ~7 t/s  
+**After initial layer optimization:** ~17 t/s  
+**After full cluster tuning:** 30–35 t/s
+
+The progression that got us there:
+
+1. **Start with `--n-gpu-layers 99`** — tells llama-server to offload as many layers as possible. With Mistral-Nemo-12B Q4_K_M this results in all 41/41 layers offloading to jarvis GPU via RPC.
+
+2. **Verify GPU is actually working** — watch the llama-server startup log for:
+   ```
+   load_tensors: offloaded 41/41 layers to GPU
+   load_tensors: RPC[192.168.50.210:50052] model buffer size = 6763.30 MiB
+   load_tensors: CPU_Mapped model buffer size = 360.00 MiB
+   ```
+   If layers aren't offloading, the RPC connection isn't established.
+
+3. **Check actual throughput** — the timings block in llama-server responses shows real t/s. Tune from there.
+
+**Current llama-server service on ultron (`/etc/systemd/system/llama-server.service`):**
+```ini
+[Unit]
+Description=Llama.cpp Server (RPC frontend — Mistral-Nemo general)
+After=network.target
+
+[Service]
+Type=simple
+User=root
+ExecStart=/root/llama.cpp/build/bin/llama-server \
+  --model /home/gramps/models/Mistral-Nemo-Instruct-2407-Q4_K_M.gguf \
+  --rpc 192.168.50.212:50052 \
+  --host 0.0.0.0 \
+  --port 8081 \
+  --n-gpu-layers 99
+Restart=on-failure
+RestartSec=5
+
+[Install]
+WantedBy=multi-user.target
+```
+
+**llama-rpc service on jarvis (`/etc/systemd/system/llama-rpc.service`):**
+```ini
+[Unit]
+Description=Llama.cpp RPC Server (GPU backend — RX 6600 XT Vulkan)
+After=network.target
+
+[Service]
+Type=simple
+User=root
+ExecStart=/root/llama.cpp/build/bin/llama-rpc-server \
+  --host 0.0.0.0 \
+  --port 50052
+Restart=on-failure
+RestartSec=5
+
+[Install]
+WantedBy=multi-user.target
+```
+
+---
+
+## Models
+
+### Current
+| Model | Location | Port | Purpose |
+|-------|----------|------|---------|
+| Mistral-Nemo-Instruct-2407-Q4_K_M | `/home/gramps/models/` on jarvis | ultron:8081 | General assistant, chat |
+| mxbai-embed-large | ultron (Docker/Ollama) | ultron:11434 | RAG embeddings |
+
+### Planned
+| Model | Size | Port | Purpose |
+|-------|------|------|---------|
+| Qwen2.5-Coder-14B-Q5_K_M | ~10GB | ultron:8082 | Code completion, pair programming |
+
+> **Note:** ultron has 16GB RAM. Only one primary inference model can be hot at a time. llama-server instances are swapped via systemd when switching between general and code models.
+
+---
+
+## RAG System
+
+jC uses **Qdrant** for vector storage and **mxbai-embed-large** (1024-dim) for embeddings.
+
+### Qdrant Collection
+- **Collection:** `jarvis_rag`
+- **Vector size:** 1024 (mxbai-embed-large output)
+- **Distance:** Cosine
+- **Score threshold:** 0.25 (filters low-relevance chunks)
+- **Chunks retrieved per query:** 3 (configurable)
+
+### RAM Ceiling
+Each vector = 4KB (1024 dims × float32). With ultron's ~4-6GB available to Qdrant after llama-server:
+- Practical ceiling: ~1–1.5M chunks before RAM becomes the bottleneck
+- Current corpus: 219 points (early stage)
+- Storage on disk: negligible against pivault's 10.83TB
+
+### What Gets Ingested
+- Code repositories (your actual codebase)
+- Pair-programming conversation history
+- Architecture decisions and working commands
+- Documentation and URLs (fetched and stripped via beautifulsoup4/httpx)
+
+---
+
+## JarvisChat Service (`/etc/systemd/system/jarvischat.service`)
+
+```ini
+[Unit]
+Description=JarvisChat - Local LLM Developer Platform
+After=network.target
+
+[Service]
+Type=simple
+User=root
+WorkingDirectory=/opt/jarvischat
+ExecStart=/opt/jarvischat/venv/bin/uvicorn app:app --host 0.0.0.0 --port 8080
+Restart=always
+RestartSec=5
+Environment=PYTHONUNBUFFERED=1
+Environment=OLLAMA_BASE=http://192.168.50.108:8081
+Environment=LLAMA_SERVER_BASE=http://192.168.50.108:8081
+
+[Install]
+WantedBy=multi-user.target
+```
+
+---
+
+## Installation
+
+### Prerequisites
+- Python 3.11+ (tested on 3.13)
+- llama.cpp built from source on both jarvis (RPC server) and ultron (llama-server)
+- Qdrant running on ultron
+- Ollama on ultron (for mxbai-embed-large embeddings)
+- SearXNG on jarvis:8888 (optional, for web search)
+
+### Fresh Install
+
+```bash
+sudo mkdir -p /opt/jarvischat
+sudo chown $USER:$USER /opt/jarvischat
+cd /opt/jarvischat
+python3 -m venv venv
+./venv/bin/pip install fastapi uvicorn httpx psutil jinja2 python-multipart qdrant-client
+mkdir -p templates static
+```
+
+Copy `app.py` to `/opt/jarvischat/` and `index.html` to `/opt/jarvischat/templates/`.
+
+### Bootstrap the PIN
+
+```bash
+export JARVISCHAT_ADMIN_PIN=XXXX  # your 4-digit PIN
+```
+
+Or allow the insecure default for testing:
+```bash
+export JARVISCHAT_ALLOW_DEFAULT_PIN=true
+```
+
+### Environment Variables
+
+| Variable | Default | Description |
+|----------|---------|-------------|
+| `OLLAMA_BASE` | `http://localhost:11434` | Ollama-compatible endpoint (legacy) |
+| `LLAMA_SERVER_BASE` | `http://192.168.50.108:8081` | llama-server OpenAI-compat inference endpoint |
+| `JARVISCHAT_ADMIN_PIN` | (none) | 4-digit admin PIN (required on first boot) |
+| `JARVISCHAT_ALLOW_DEFAULT_PIN` | `false` | Allow insecure default PIN 1234 |
+| `JARVISCHAT_TRUSTED_ORIGINS` | (none) | Comma-separated trusted origins for CSRF |
+| `JARVISCHAT_ALLOWED_CIDRS` | RFC1918 + loopback | Allowed client IP CIDRs |
+
+---
+
+## API Endpoints
+
+### Auth
+| Method | Path | Description |
+|--------|------|-------------|
+| POST | `/api/auth/guest` | Create guest session |
+| POST | `/api/auth/login` | Admin PIN login |
+| POST | `/api/auth/logout` | Revoke session |
+| GET | `/api/auth/session` | Check session status |
+| POST | `/api/auth/heartbeat` | Keep session alive |
+
+### Chat & Search
+| Method | Path | Description |
+|--------|------|-------------|
+| POST | `/api/chat` | Streaming chat (SSE) |
+| POST | `/api/search` | Explicit web search via SearXNG |
+| GET | `/api/search/status` | SearXNG health check |
+
+### Models
+| Method | Path | Description |
+|--------|------|-------------|
+| GET | `/api/models` | List available models from llama-server |
+| GET | `/api/ps` | Running models |
+| POST | `/api/show` | Model info |
+
+### Memory
+| Method | Path | Description |
+|--------|------|-------------|
+| GET | `/api/memories` | List all memories |
+| POST | `/api/memories` | Add memory |
+| PUT | `/api/memories/{rowid}` | Update memory |
+| DELETE | `/api/memories/{rowid}` | Delete memory |
+| GET | `/api/memories/search?q=` | FTS5 search memories |
+| GET | `/api/memories/stats` | Memory statistics |
+
+### Conversations
+| Method | Path | Description |
+|--------|------|-------------|
+| GET | `/api/conversations` | List conversations |
+| POST | `/api/conversations` | Create conversation |
+| GET | `/api/conversations/{id}` | Get conversation + messages |
+| PUT | `/api/conversations/{id}` | Update title/model |
+| DELETE | `/api/conversations/{id}` | Delete conversation |
+| DELETE | `/api/conversations` | Delete all conversations |
+
+### Profile & Settings
+| Method | Path | Description |
+|--------|------|-------------|
+| GET | `/api/profile` | Get profile |
+| PUT | `/api/profile` | Update profile |
+| GET | `/api/settings` | Get settings |
+| PUT | `/api/settings` | Update settings |
+| GET | `/api/stats` | CPU/RAM/GPU stats |
+
+### Skills
+| Method | Path | Description |
+|--------|------|-------------|
+| GET | `/api/skills` | List all skills |
+| GET | `/api/skills/active` | List enabled skills |
+| PUT | `/api/skills/{key}` | Enable/disable skill |
+
+---
+
+## Memory Commands
+
+Say these in chat to interact with the memory system:
+
+| Command | Effect |
+|---------|--------|
+| `remember that [fact]` | Stores fact in FTS5 memory |
+| `please remember [fact]` | Same |
+| `don't forget [fact]` | Same |
+| `forget about [topic]` | Deletes matching memories |
+
+---
+
+## Troubleshooting
+
+### jC starts but inference is slow or failing
+Check that llama-rpc is running on jarvis and llama-server is connected:
+```bash
+# On jarvis
+systemctl status llama-rpc
+
+# On ultron — look for "offloaded N/N layers to GPU" in logs
+journalctl -u llama-server -n 50 --no-pager
+```
+
+### ultron shows no CPU activity during inference
+Inference is being handled entirely by jarvis GPU via RPC — this is correct and expected. ultron's CPU is only involved for non-offloaded tensors (a small fraction of the model).
+
+### RAG not returning results
+Check Qdrant is up and the collection exists:
+```bash
+curl http://192.168.50.108:6333/collections/jarvis_rag
+```
+Verify `points_count` > 0. If zero, the corpus hasn't been seeded yet.
+
+### jC won't start — PIN bootstrap error
+Set the PIN via environment before first boot:
+```bash
+export JARVISCHAT_ADMIN_PIN=XXXX
+systemctl restart jarvischat
+```
+
+### sqlite3 not found
+Use Python instead:
+```bash
+python3 -c "import sqlite3; print(sqlite3.connect('/opt/jarvischat/jarvischat.db').execute('SELECT * FROM settings').fetchall())"
+```
+
+---
+
+## Roadmap
+
+### TODO (Priority Order)
+1. **Tool calling** — read_file/write_file with /opt/jarvischat whitelist, tool_calls dispatch loop
+2. **git_tool** — Gitea integration for commit/push from jC
+3. **Audit logging** — structured audit trail to syslog
+4. SearXNG persistence (DONE ✅)
+5. search+ prefix for explicit search
+6. profile.example.md
+7. Conversation search/filter
+8. Export to markdown
+9. Keyboard shortcuts
+10. Retry button
+11. Source links in responses
+12. Rename conversations
+13. Multiple profiles
+14. KWIC auto-tags
+15. Image input (vision)
+16. btop split-screen integration
+17. Containerize
+18. SearXNG health indicator in UI
+19. check_patch_notes tool
+20. GitLab mirror of llgit repo
+
+### ROADMAP (Longer Horizon)
+
+**(A) Modular refactor** — Split monolithic app.py into routers/, services/, config.py, db.py, auth.py. Prerequisite for everything below.
+
+**(B) RAG ingest/manage UI** — File upload, URL ingest (fetch + strip HTML via beautifulsoup4/httpx, store URL as source metadata for citation), delete chunks/collections.
+
+**(C) Backend config panel** — Switch between Ollama/llama-server, endpoint URLs, model switching, restart — all from the UI without touching config files.
+
+**(D) Response metrics display** — tokens/sec, TTFT, context size, RAG chunks retrieved + scores — visible in the UI per response.
+
+**(E) Response quality feedback** — thumbs/stars/tags per response → feedback corpus → future RLHF dataset.
+
+**(F) IDE integration** — Continue.dev + VS Code, pointed at jC:8080 (not direct to inference endpoint). All IDE traffic — including pair-programming conversations — goes through jC so sessions are persisted and become RAG-worthy content. jC needs FIM request format handling to support inline autocomplete.
+
+**(G) Conversation history export → RAG ingest** — Bulk ingest existing conversation history into Qdrant.
+
+**(H) Fine-tuning pipeline** — LoRA on Mistral-Nemo from feedback corpus (item E).
+
+**(I) Autonomous RAG** — At conversation end, jC self-evaluates the transcript, extracts significant chunks (solved problems, working commands, architectural decisions), and ingests them into Qdrant automatically with metadata (date, conversation_id, reason). jC decides what it needs to remember. Closes the loop.
+
+**(J) Startup hardware/resource self-assessment** — On boot, jC queries ultron for available RAM, Qdrant consumption, and llama-server footprint. Derives dynamic high-water marks for RAG chunk limits, context window sizing, retrieval limits, and eviction thresholds. Writes a living config file. Replaces magic numbers with runtime-negotiated values.
+
+**(K) RAG corpus management** — Weighted LRU eviction with composite score (recency + frequency + content age) + manual pin flag for load-bearing knowledge. Prevents corpus bloat from degrading retrieval quality. Analogous to memcache eviction policy.
+
+**(L) Dual inference model architecture** — Mistral-Nemo-12B on ultron:8081 (general assistant), Qwen2.5-Coder-14B-Q5_K_M on ultron:8082 (code/pair programming). jC selects endpoint based on active model. Only one model hot at a time given ultron's 16GB RAM constraint.
+
+---
+
+## Primary Cluster Objectives
+
+1. **Generative AI inference** — Local, private, fast enough to be useful
+2. **Agentic functionality** — Autonomous RAG self-management is the canonical first example. The system acts, not just responds.
+
+---
+
+## Repository
+
+```
+ssh://gitea@llgit.llamachile.tube:1319/gramps/jarvisChat.git
+```
+
+> SSH username is `gitea`, not `git`. Port 1319.
+
+---
+
+## License
+
+MIT

app.py.bak

@@ -1,6 +1,6 @@
 #!/usr/bin/env python3
 """
-JarvisChat - Lightweight Ollama Coding Companion
+JarvisChat - Privacy-First Homelab Developer Platform
 A minimal replacement for Open-WebUI that actually runs on Python 3.13
 Talks to Ollama API on localhost:11434
 
@@ -56,8 +56,9 @@ syslog_handler.setFormatter(
 log.addHandler(syslog_handler)
 
 # --- Configuration ---
-VERSION = "1.7.6"
+VERSION = "v1.8.0"
 OLLAMA_BASE = os.environ.get("OLLAMA_BASE", "http://localhost:11434")
+LLAMA_SERVER_BASE = os.environ.get("LLAMA_SERVER_BASE", "http://192.168.50.108:8081")
 SEARXNG_BASE = "http://localhost:8888"
 BASE_DIR = Path(__file__).parent
 DB_PATH = BASE_DIR / "jarvischat.db"
@@ -1038,7 +1039,7 @@ def get_gpu_stats() -> dict:
 @asynccontextmanager
 async def lifespan(app: FastAPI):
     log.info(f"JarvisChat v{VERSION} starting up")
-    log.info(f"Ollama: {OLLAMA_BASE}, SearXNG: {SEARXNG_BASE}")
+    log.info(f"Ollama: {OLLAMA_BASE}, llama-server: {LLAMA_SERVER_BASE}, SearXNG: {SEARXNG_BASE}")
     init_db()
     log.info(f"Memory system: {get_memory_count()} memories loaded")
     yield
@@ -1966,22 +1967,18 @@ async def explicit_search(request: Request):
             try:
                 async with client.stream(
                     "POST",
-                    f"{OLLAMA_BASE}/api/chat",
+                    f"{LLAMA_SERVER_BASE}/v1/chat/completions",
                     json={"model": model, "messages": messages, "stream": True},
                     timeout=httpx.Timeout(300.0, connect=10.0),
                 ) as resp:
                     async for line in resp.aiter_lines():
                         if line.strip():
-                            try:
-                                chunk = json.loads(line)
-                                if "message" in chunk and "content" in chunk["message"]:
-                                    token = chunk["message"]["content"]
+                            token, done, _ = parse_llama_stream_chunk(line)
+                            if token:
                                 full_response.append(token)
                                 yield f"data: {json.dumps({'token': token, 'conversation_id': conv_id})}\n\n"
-                                if chunk.get("done"):
+                            if done:
                                 break
-                            except json.JSONDecodeError:
-                                pass
             except Exception as e:
                 incident_key = log_incident(
                     "search_summarization_stream",
@@ -2020,7 +2017,32 @@ async def explicit_search(request: Request):
 # =============================================================================
 
 
-def build_system_prompt(db, extra_prompt="", user_message=""):
+
+async def query_rag(query: str, limit: int = 3) -> list[dict]:
+    """Query Qdrant for semantically relevant chunks."""
+    try:
+        async with httpx.AsyncClient() as client:
+            embed_resp = await client.post(
+                "http://192.168.50.108:11434/api/embeddings",
+                json={"model": "mxbai-embed-large", "prompt": query},
+                timeout=10.0,
+            )
+            if embed_resp.status_code != 200:
+                return []
+            vector = embed_resp.json()["embedding"]
+            search_resp = await client.post(
+                "http://192.168.50.108:6333/collections/jarvis_rag/points/search",
+                json={"vector": vector, "limit": limit, "with_payload": True},
+                timeout=10.0,
+            )
+            if search_resp.status_code != 200:
+                return []
+            return search_resp.json().get("result", [])
+    except Exception as e:
+        log.warning(f"RAG query error: {e}")
+        return []
+
+async def build_system_prompt(db, extra_prompt="", user_message=""):
     """Build the full system prompt: profile + memories + preset."""
     parts = []
     settings = {
@@ -2040,6 +2062,17 @@ def build_system_prompt(db, extra_prompt="", user_message=""):
             parts.append("## Relevant Context from Memory\n" + "\n".join(memory_lines))
             log.debug(f"Injected {len(memories)} memories into context")
 
+    if user_message:
+        try:
+            rag_results = await query_rag(user_message)
+            if rag_results:
+                rag_lines = [r["payload"]["text"] for r in rag_results if r["score"] > 0.25]
+                if rag_lines:
+                    parts.append("## Retrieved Context\n" + "\n\n---\n\n".join(rag_lines))
+                    log.warning(f"RAG injected {len(rag_lines)} chunks into context")
+        except Exception as e:
+            log.warning(f"RAG injection error: {e}")
+
     if settings.get("skills_enabled", "true") == "true":
         active_skills = [s for s in list_skills_with_state(db) if s["enabled"]]
         if active_skills:
@@ -2051,6 +2084,42 @@ def build_system_prompt(db, extra_prompt="", user_message=""):
     return "\n\n---\n\n".join(parts) if parts else ""
 
 
+def parse_llama_stream_chunk(line: str) -> tuple[str | None, bool, dict]:
+    """Parse OpenAI-compatible SSE chunk. Returns (token, is_done, stats)."""
+    if line.startswith("data: "):
+        line = line[6:]
+    if line.strip() == "[DONE]":
+        return None, True, {}
+    try:
+        chunk = json.loads(line)
+        # OpenAI format
+        choices = chunk.get("choices", [])
+        if choices:
+            delta = choices[0].get("delta", {})
+            token = delta.get("content")
+            finish = choices[0].get("finish_reason")
+            stats = {}
+            if finish == "stop":
+                usage = chunk.get("usage", {})
+                stats["tokens_per_sec"] = usage.get("tokens_per_second", 0.0)
+            return token, finish == "stop", stats
+        # Ollama format fallback
+        if "message" in chunk and "content" in chunk["message"]:
+            token = chunk["message"]["content"]
+            done = chunk.get("done", False)
+            stats = {}
+            if done:
+                eval_count = chunk.get("eval_count", 0)
+                eval_duration = chunk.get("eval_duration", 0)
+                stats["tokens_per_sec"] = (
+                    (eval_count / (eval_duration / 1e9)) if eval_duration > 0 else 0
+                )
+            return token, done, stats
+    except json.JSONDecodeError:
+        pass
+    return None, False, {}
+
+
 @app.post("/api/chat")
 async def chat(request: Request):
     body = await read_json_body(request, BODY_LIMIT_CHAT_BYTES)
@@ -2096,7 +2165,7 @@ async def chat(request: Request):
         "SELECT role, content FROM messages WHERE conversation_id = ? ORDER BY id ASC",
         (conv_id,),
     ).fetchall()
-    system_prompt = build_system_prompt(db, preset_prompt, user_message)
+    system_prompt = await build_system_prompt(db, preset_prompt, user_message)
     db.close()
 
     messages = []
@@ -2109,7 +2178,6 @@ async def chat(request: Request):
         "model": model,
         "messages": messages,
         "stream": True,
-        "logprobs": True,
     }
 
     async def stream_response():
@@ -2124,31 +2192,18 @@ async def chat(request: Request):
             try:
                 async with client.stream(
                     "POST",
-                    f"{OLLAMA_BASE}/api/chat",
+                    f"{LLAMA_SERVER_BASE}/v1/chat/completions",
                     json=ollama_payload,
                     timeout=httpx.Timeout(300.0, connect=10.0),
                 ) as resp:
                     async for line in resp.aiter_lines():
                         if line.strip():
-                            try:
-                                chunk = json.loads(line)
-                                if "message" in chunk and "content" in chunk["message"]:
-                                    token = chunk["message"]["content"]
+                            token, done, stats = parse_llama_stream_chunk(line)
+                            if token:
                                 full_response.append(token)
                                 yield f"data: {json.dumps({'token': token, 'conversation_id': conv_id})}\n\n"
-                                if "logprobs" in chunk and chunk["logprobs"]:
-                                    all_logprobs.extend(chunk["logprobs"])
-                                if chunk.get("done"):
-                                    eval_count = chunk.get("eval_count", 0)
-                                    eval_duration = chunk.get("eval_duration", 0)
-                                    tokens_per_sec = (
-                                        (eval_count / (eval_duration / 1e9))
-                                        if eval_duration > 0
-                                        else 0
-                                    )
-                                    break
-                            except json.JSONDecodeError:
-                                pass
+                            if done:
+                                tokens_per_sec = stats.get("tokens_per_sec", 0.0)
 
                 assistant_msg = "".join(full_response)
                 perplexity = calculate_perplexity(all_logprobs) if all_logprobs else 0.0
@@ -2186,7 +2241,7 @@ async def chat(request: Request):
                         augmented_response = []
                         async with client.stream(
                             "POST",
-                            f"{OLLAMA_BASE}/api/chat",
+                            f"{LLAMA_SERVER_BASE}/v1/chat/completions",
                             json={
                                 "model": model,
                                 "messages": augmented_messages,
@@ -2196,19 +2251,11 @@ async def chat(request: Request):
                         ) as resp2:
                             async for line in resp2.aiter_lines():
                                 if line.strip():
-                                    try:
-                                        chunk = json.loads(line)
-                                        if (
-                                            "message" in chunk
-                                            and "content" in chunk["message"]
-                                        ):
-                                            augmented_response.append(
-                                                chunk["message"]["content"]
-                                            )
-                                        if chunk.get("done"):
+                                    token2, done2, _ = parse_llama_stream_chunk(line)
+                                    if token2:
+                                        augmented_response.append(token2)
+                                    if done2:
                                         break
-                                    except json.JSONDecodeError:
-                                        pass
 
                         raw_response = "".join(augmented_response) or assistant_msg
                         cleaned_response = clean_hedging(raw_response)
@@ -2261,6 +2308,8 @@ async def chat(request: Request):
 
                 yield f"data: {json.dumps({'done': True, 'conversation_id': conv_id, 'perplexity': round(perplexity, 2), 'tokens_per_sec': round(tokens_per_sec, 1)})}\n\n"
 
+            except httpx.RemoteProtocolError:
+                pass  # llama-server closes connection after [DONE] — normal
             except httpx.ConnectError:
                 yield f"data: {json.dumps({'error': 'Cannot connect to Ollama. Is it running?'})}\n\n"
             except Exception as e:

auth.py

@@ -0,0 +1,204 @@
+"""
+JarvisChat - Auth: session management, PIN verification, middleware, auth routes.
+"""
+import hashlib
+import hmac
+import logging
+import re
+import time
+import uuid
+from typing import Optional
+
+from fastapi import APIRouter, HTTPException, Request
+from fastapi.responses import JSONResponse
+
+from config import SESSION_TIMEOUT_SECONDS, MAX_PIN_ATTEMPTS, PIN_LOCKOUT_SECONDS, RATE_WINDOW_SECONDS
+from db import get_db, get_setting
+from security import (
+    SESSIONS, PIN_ATTEMPTS, SESSION_LOCK, audit_event, get_client_ip,
+    is_ip_allowed, check_rate_limit, rate_policy, origin_allowed,
+    is_state_changing, request_body_limit, read_json_body, hash_pin,
+    customer_error_envelope, log_incident,
+)
+
+log = logging.getLogger("jarvischat")
+router = APIRouter()
+
+
+def verify_admin_pin(pin: str) -> bool:
+    if not re.fullmatch(r"\d{4}", pin or ""):
+        return False
+    db = get_db()
+    pin_hash = get_setting(db, "admin_pin_hash", "")
+    pin_salt = get_setting(db, "admin_pin_salt", "")
+    db.close()
+    if not pin_hash or not pin_salt:
+        return False
+    _, candidate_hash = hash_pin(pin, salt_hex=pin_salt)
+    return hmac.compare_digest(candidate_hash, pin_hash)
+
+
+def is_ip_locked(ip: str) -> tuple:
+    now_ts = time.time()
+    with SESSION_LOCK:
+        state = PIN_ATTEMPTS.get(ip)
+        if not state:
+            return False, 0
+        locked_until = state.get("locked_until", 0)
+        if locked_until > now_ts:
+            return True, int(locked_until - now_ts)
+        if locked_until:
+            PIN_ATTEMPTS.pop(ip, None)
+    return False, 0
+
+
+def record_pin_failure(ip: str) -> None:
+    now_ts = time.time()
+    with SESSION_LOCK:
+        state = PIN_ATTEMPTS.get(ip, {"fail_count": 0, "locked_until": 0})
+        state["fail_count"] = int(state.get("fail_count", 0)) + 1
+        if state["fail_count"] >= MAX_PIN_ATTEMPTS:
+            state["locked_until"] = now_ts + PIN_LOCKOUT_SECONDS
+            state["fail_count"] = 0
+        PIN_ATTEMPTS[ip] = state
+
+
+def clear_pin_failures(ip: str) -> None:
+    with SESSION_LOCK:
+        PIN_ATTEMPTS.pop(ip, None)
+
+
+def cleanup_sessions(now_ts: Optional[float] = None) -> None:
+    now_ts = now_ts or time.time()
+    with SESSION_LOCK:
+        expired = [
+            sid for sid, meta in SESSIONS.items()
+            if (now_ts - meta.get("last_seen", 0)) > SESSION_TIMEOUT_SECONDS
+        ]
+        for sid in expired:
+            del SESSIONS[sid]
+
+
+def create_session(ip: str, role: str) -> str:
+    now_ts = time.time()
+    sid = uuid.uuid4().hex
+    with SESSION_LOCK:
+        SESSIONS[sid] = {"ip": ip, "role": role, "created_at": now_ts, "last_seen": now_ts}
+    return sid
+
+
+def validate_session(sid: str, ip: str, touch: bool = True) -> bool:
+    if not sid:
+        return False
+    now_ts = time.time()
+    cleanup_sessions(now_ts)
+    with SESSION_LOCK:
+        session = SESSIONS.get(sid)
+        if not session or session.get("ip") != ip:
+            return False
+        if touch:
+            session["last_seen"] = now_ts
+    return True
+
+
+def get_session(sid: str, ip: str, touch: bool = True) -> Optional[dict]:
+    if not sid:
+        return None
+    now_ts = time.time()
+    cleanup_sessions(now_ts)
+    with SESSION_LOCK:
+        session = SESSIONS.get(sid)
+        if not session or session.get("ip") != ip:
+            return None
+        if touch:
+            session["last_seen"] = now_ts
+        return dict(session)
+
+
+def revoke_session(sid: str) -> None:
+    if not sid:
+        return
+    with SESSION_LOCK:
+        SESSIONS.pop(sid, None)
+
+
+def is_admin_only(path: str, method: str) -> bool:
+    if method in {"PUT", "DELETE", "PATCH"}:
+        return True
+    if method != "POST":
+        return False
+    guest_allowed_posts = {
+        "/api/chat", "/api/search", "/api/show", "/api/auth/login",
+        "/api/auth/logout", "/api/auth/session", "/api/auth/heartbeat", "/api/auth/guest",
+    }
+    return path not in guest_allowed_posts
+
+
+# --- Auth routes ---
+
+@router.post("/api/auth/guest")
+async def auth_guest(request: Request):
+    ip = get_client_ip(request)
+    sid = create_session(ip, role="guest")
+    audit_event("guest_session", "success", ip=ip, role="guest")
+    return {"status": "ok", "session_id": sid, "role": "guest", "timeout_seconds": SESSION_TIMEOUT_SECONDS}
+
+
+@router.post("/api/auth/login")
+async def auth_login(request: Request):
+    from security import BODY_LIMIT_DEFAULT_BYTES
+    body = await read_json_body(request, BODY_LIMIT_DEFAULT_BYTES)
+    pin = str(body.get("pin", ""))
+    ip = get_client_ip(request)
+    locked, retry_after = is_ip_locked(ip)
+    if locked:
+        audit_event("admin_login", "locked", ip=ip, role="none", details=f"retry_after={retry_after}", warning=True)
+        raise HTTPException(status_code=429, detail=f"Too many failed PIN attempts. Retry in {retry_after}s.")
+    if not verify_admin_pin(pin):
+        record_pin_failure(ip)
+        audit_event("admin_login", "failed", ip=ip, role="none", warning=True)
+        raise HTTPException(status_code=401, detail="Invalid PIN")
+    clear_pin_failures(ip)
+    sid = create_session(ip, role="admin")
+    audit_event("admin_login", "success", ip=ip, role="admin")
+    return {"status": "ok", "session_id": sid, "role": "admin", "timeout_seconds": SESSION_TIMEOUT_SECONDS}
+
+
+@router.get("/api/auth/session")
+async def auth_session(request: Request):
+    sid = request.headers.get("x-session-id", "").strip()
+    ip = get_client_ip(request)
+    session = get_session(sid, ip, touch=True)
+    return {"authenticated": bool(session), "role": session.get("role") if session else "none"}
+
+
+@router.post("/api/auth/heartbeat")
+async def auth_heartbeat(request: Request):
+    sid = request.headers.get("x-session-id", "").strip()
+    ip = get_client_ip(request)
+    if not sid or not validate_session(sid, ip, touch=True):
+        raise HTTPException(status_code=401, detail="Authentication required")
+    return {"status": "ok"}
+
+
+@router.post("/api/auth/logout")
+async def auth_logout(request: Request):
+    from security import BODY_LIMIT_DEFAULT_BYTES
+    ip = get_client_ip(request)
+    sid = request.headers.get("x-session-id", "").strip()
+    role = "none"
+    if sid:
+        session = get_session(sid, ip, touch=False)
+        role = session.get("role", "none") if session else "none"
+    if not sid:
+        try:
+            body = await read_json_body(request, BODY_LIMIT_DEFAULT_BYTES)
+            sid = str(body.get("session_id", "")).strip()
+        except Exception:
+            try:
+                sid = (await request.body()).decode("utf-8", errors="ignore").strip()
+            except Exception:
+                sid = ""
+    revoke_session(sid)
+    audit_event("logout", "success", ip=ip, role=role)
+    return {"status": "ok"}

config.py

@@ -0,0 +1,159 @@
+"""
+JarvisChat - Central configuration.
+All constants, environment variables, limits, and skill registry live here.
+"""
+import os
+import re
+import ipaddress
+import logging
+
+log = logging.getLogger("jarvischat")
+
+VERSION = "v1.8.0"
+OLLAMA_BASE = os.environ.get("OLLAMA_BASE", "http://localhost:11434")
+LLAMA_SERVER_BASE = os.environ.get("LLAMA_SERVER_BASE", "http://192.168.50.108:8081")
+SEARXNG_BASE = "http://localhost:8888"
+DEFAULT_MODEL = "llama3.1:latest"
+
+# --- Auth ---
+SESSION_TIMEOUT_SECONDS = 90
+MAX_PIN_ATTEMPTS = 5
+PIN_LOCKOUT_SECONDS = 300
+ALLOW_DEFAULT_PIN = os.getenv("JARVISCHAT_ALLOW_DEFAULT_PIN", "false").lower() == "true"
+TRUSTED_ORIGINS = {
+    origin.strip().rstrip("/")
+    for origin in os.getenv("JARVISCHAT_TRUSTED_ORIGINS", "").split(",")
+    if origin.strip()
+}
+DEFAULT_ALLOWED_CIDRS = "127.0.0.0/8,::1/128,10.0.0.0/8,172.16.0.0/12,192.168.0.0/16"
+ALLOWED_CIDRS_RAW = os.getenv("JARVISCHAT_ALLOWED_CIDRS", DEFAULT_ALLOWED_CIDRS)
+TRUST_X_FORWARDED_FOR = (
+    os.getenv("JARVISCHAT_TRUST_X_FORWARDED_FOR", "false").lower() == "true"
+)
+
+# --- Rate limits ---
+RATE_WINDOW_SECONDS = 60
+RL_LOGIN_PER_WINDOW = 10
+RL_CHAT_PER_WINDOW = 24
+RL_SEARCH_PER_WINDOW = 16
+RL_WRITE_PER_WINDOW = 30
+RL_DEFAULT_PER_WINDOW = 240
+RL_STATS_PER_WINDOW = 600
+
+# --- Payload limits ---
+BODY_LIMIT_DEFAULT_BYTES = 64 * 1024
+BODY_LIMIT_CHAT_BYTES = 128 * 1024
+BODY_LIMIT_PROFILE_BYTES = 256 * 1024
+
+MAX_CHAT_MESSAGE_CHARS = 8000
+MAX_SEARCH_QUERY_CHARS = 500
+MAX_PROFILE_CHARS = 32000
+MAX_MEMORY_FACT_CHARS = 2000
+MAX_PRESET_NAME_CHARS = 120
+MAX_PRESET_PROMPT_CHARS = 12000
+MAX_SETTINGS_KEYS = 16
+MAX_SETTINGS_VALUE_CHARS = 8000
+MAX_CONVERSATION_TITLE_CHARS = 200
+MAX_SKILL_KEY_CHARS = 120
+MAX_SKILL_PROMPT_CHARS = 1600
+
+ALLOWED_SETTINGS_KEYS = {
+    "profile_enabled",
+    "default_model",
+    "search_enabled",
+    "memory_enabled",
+    "skills_enabled",
+}
+
+# --- Perplexity ---
+PERPLEXITY_THRESHOLD = 15.0
+
+# --- Refusal / hedge patterns ---
+REFUSAL_PATTERNS = re.compile(
+    r"|".join([
+        r"i don'?t have (?:real-?time|current|live)",
+        r"i (?:can'?t|cannot) provide (?:current|real-?time|live)",
+        r"i don'?t have access to (?:current|real-?time|live)",
+        r"(?:current|live|real-?time) (?:data|information|prices?|weather)",
+        r"my (?:knowledge|training) (?:cutoff|only goes|ends)",
+        r"as of my (?:knowledge|training) cutoff",
+        r"i'?m not able to (?:access|provide|browse)",
+        r"(?:check|visit|use) a (?:website|financial|news)",
+        r"as an ai model",
+        r"based on my training data",
+        r"i don'?t have the capability",
+    ]),
+    re.IGNORECASE,
+)
+
+HEDGE_PATTERNS = [
+    r"^I'?m sorry,?\s*but\s*I\s*(?:can'?t|cannot)\s*assist\s*with\s*that[^.]*\.\s*",
+    r"^I'?m sorry,?\s*but[^.]*(?:previous|incorrect)[^.]*\.\s*",
+    r"(?:But\s+)?[Pp]lease\s+(?:make\s+sure\s+to\s+)?verify\s+(?:the\s+)?(?:data|information|this)\s+(?:from\s+)?(?:reliable\s+)?sources[^.]*\.\s*",
+    r"[Pp]lease\s+verify[^.]*(?:accurate|reliability)[^.]*\.\s*",
+    r"[Bb]ut\s+please\s+(?:make\s+sure|verify|check)[^.]*\.\s*",
+]
+
+# --- Built-in skills registry ---
+BUILTIN_SKILLS = [
+    {"key": "memory.search",       "name": "Memory Search",        "category": "memory",       "risk": "low",    "description": "Search stored memory facts relevant to the current prompt."},
+    {"key": "memory.add",          "name": "Memory Add",           "category": "memory",       "risk": "medium", "description": "Store a new memory fact with topic tagging."},
+    {"key": "memory.forget",       "name": "Memory Forget",        "category": "memory",       "risk": "high",   "description": "Delete matching memories when asked to forget information."},
+    {"key": "conversation.list",   "name": "Conversation List",    "category": "conversation", "risk": "low",    "description": "List existing conversations with metadata."},
+    {"key": "conversation.get",    "name": "Conversation Get",     "category": "conversation", "risk": "low",    "description": "Read a conversation and its message history."},
+    {"key": "conversation.delete", "name": "Conversation Delete",  "category": "conversation", "risk": "high",   "description": "Delete a single conversation thread."},
+    {"key": "conversation.delete_all", "name": "Conversation Delete All", "category": "conversation", "risk": "high", "description": "Delete all conversations and messages."},
+    {"key": "search.web",          "name": "Web Search",           "category": "search",       "risk": "low",    "description": "Run explicit web search and summarize results."},
+    {"key": "settings.get",        "name": "Settings Get",         "category": "settings",     "risk": "low",    "description": "Read current runtime settings."},
+    {"key": "settings.update",     "name": "Settings Update",      "category": "settings",     "risk": "high",   "description": "Update allowlisted runtime settings keys."},
+]
+
+SKILLS_BY_KEY = {s["key"]: s for s in BUILTIN_SKILLS}
+
+
+def parse_allowed_cidrs(raw: str) -> list:
+    networks = []
+    for entry in (raw or "").split(","):
+        value = entry.strip()
+        if not value:
+            continue
+        try:
+            networks.append(ipaddress.ip_network(value, strict=False))
+        except ValueError:
+            log.warning(f"Invalid CIDR ignored: {value}")
+    return networks
+
+
+ALLOWED_NETWORKS = parse_allowed_cidrs(ALLOWED_CIDRS_RAW)
+
+DEFAULT_PROFILE = """You are a coding companion running locally on a machine called "jarvis".
+
+## Environment
+- jarvis: Debian 13 (trixie) x86_64, AMD Ryzen 5 5600X, 16GB RAM, AMD RX 6600 XT (8GB VRAM)
+- ultron: Debian 13, Ryzen 7 7840HS, 16GB RAM, primary AI inference node, IP 192.168.50.108
+- Corsair: Windows 11, gaming/streaming rig, RTX 5070 Ti
+- pivault: RPi 5, 8GB RAM, Debian 13, 11TB RAID5 NAS at /mnt/pivault, IP 192.168.50.158
+- Router: ASUS ROG Rapture GT-BE98 Pro "BigBlinkyRouter" at 192.168.50.1
+- llama-server on ultron:8081 (OpenAI-compat API), Qdrant on ultron:6333
+
+## About the User
+- Experienced developer, BS in Computer Science (Oklahoma State), coding since 1981 (TRS-80)
+- Deep Unix/Linux background — wrote device drivers at SCO during Xenix era (1990s)
+- Currently learning Rust, transitioning from decades of PHP
+- Building a WW2 mobile game in Godot Engine for Android
+- Veteran on fixed income — prefers free/open-source solutions
+- Home lab enthusiast with Zigbee, Z-Wave and Tapo smart home devices
+
+## How to Respond
+- Be direct and concise — no hand-holding, this user knows what they're doing
+- When showing code, prefer complete working examples over snippets
+- Default to command-line solutions over GUI when possible
+- Consider resource constraints (fixed income, specific hardware limits)
+- Use Rust, Python, or bash unless another language is specifically needed
+- Explain trade-offs when multiple approaches exist"""
+
+DEFAULT_PRESETS = [
+    {"name": "Coding Companion", "prompt": "You are a senior software engineer and coding companion. Focus on writing clean, efficient, well-documented code. Provide complete working examples. Explain architectural decisions and trade-offs. Prefer Rust, Python, and bash."},
+    {"name": "Linux Sysadmin",   "prompt": "You are an experienced Linux systems administrator. Focus on command-line solutions, systemd services, networking, storage, and security. Prefer Debian/Ubuntu conventions. Be concise and direct."},
+    {"name": "General Assistant","prompt": "You are a helpful general-purpose assistant. Be clear and concise."},
+]

db.py

@@ -0,0 +1,160 @@
+"""
+JarvisChat - Database layer.
+Schema init, connection factory, settings helpers, skill state management.
+"""
+import logging
+import os
+import re
+import sqlite3
+import uuid
+from datetime import datetime, timezone
+from pathlib import Path
+from typing import Optional
+
+from config import (
+    BUILTIN_SKILLS, DEFAULT_MODEL, DEFAULT_PRESETS, DEFAULT_PROFILE,
+    MAX_SKILL_PROMPT_CHARS, ALLOWED_NETWORKS,
+)
+
+log = logging.getLogger("jarvischat")
+
+BASE_DIR = Path(__file__).parent
+DB_PATH = BASE_DIR / "jarvischat.db"
+
+
+def get_db():
+    conn = sqlite3.connect(DB_PATH)
+    conn.row_factory = sqlite3.Row
+    conn.execute("PRAGMA foreign_keys = ON")
+    return conn
+
+
+def get_setting(db, key: str, default: str = "") -> str:
+    row = db.execute("SELECT value FROM settings WHERE key = ?", (key,)).fetchone()
+    return row["value"] if row else default
+
+
+def list_skills_with_state(db) -> list:
+    rows = db.execute("SELECT skill_key, enabled, updated_at FROM skills").fetchall()
+    state_by_key = {
+        row["skill_key"]: {"enabled": bool(row["enabled"]), "updated_at": row["updated_at"]}
+        for row in rows
+    }
+    merged = []
+    for skill in BUILTIN_SKILLS:
+        state = state_by_key.get(skill["key"], {"enabled": True, "updated_at": ""})
+        merged.append({**skill, "enabled": state["enabled"], "updated_at": state["updated_at"]})
+    return sorted(merged, key=lambda s: (s["category"], s["name"]))
+
+
+def set_skill_enabled(db, skill_key: str, enabled: bool) -> None:
+    now = datetime.now(timezone.utc).isoformat()
+    db.execute(
+        "INSERT OR REPLACE INTO skills (skill_key, enabled, updated_at) VALUES (?, ?, ?)",
+        (skill_key, 1 if enabled else 0, now),
+    )
+
+
+def format_active_skills_prompt(skills: list) -> str:
+    lines = [
+        "## Active Skills",
+        "Use these skills only when needed. Prefer concise answers over unnecessary tool usage.",
+    ]
+    for skill in skills:
+        lines.append(f"- {skill['key']} ({skill['risk']} risk): {skill['description']}")
+    text = "\n".join(lines)
+    if len(text) > MAX_SKILL_PROMPT_CHARS:
+        return text[:MAX_SKILL_PROMPT_CHARS - 3] + "..."
+    return text
+
+
+def init_db():
+    from security import hash_pin
+    conn = sqlite3.connect(DB_PATH)
+    conn.row_factory = sqlite3.Row
+
+    conn.execute("""
+        CREATE TABLE IF NOT EXISTS conversations (
+            id TEXT PRIMARY KEY, title TEXT NOT NULL DEFAULT 'New Chat',
+            model TEXT NOT NULL, created_at TEXT NOT NULL, updated_at TEXT NOT NULL
+        )
+    """)
+    conn.execute("""
+        CREATE TABLE IF NOT EXISTS messages (
+            id INTEGER PRIMARY KEY AUTOINCREMENT, conversation_id TEXT NOT NULL,
+            role TEXT NOT NULL, content TEXT NOT NULL, created_at TEXT NOT NULL,
+            FOREIGN KEY (conversation_id) REFERENCES conversations(id) ON DELETE CASCADE
+        )
+    """)
+    conn.execute("""
+        CREATE TABLE IF NOT EXISTS system_presets (
+            id TEXT PRIMARY KEY, name TEXT NOT NULL, prompt TEXT NOT NULL,
+            is_default INTEGER NOT NULL DEFAULT 0, created_at TEXT NOT NULL
+        )
+    """)
+    conn.execute("""
+        CREATE TABLE IF NOT EXISTS profile (
+            id INTEGER PRIMARY KEY CHECK (id = 1), content TEXT NOT NULL, updated_at TEXT NOT NULL
+        )
+    """)
+    conn.execute("CREATE TABLE IF NOT EXISTS settings (key TEXT PRIMARY KEY, value TEXT NOT NULL)")
+    conn.execute("""
+        CREATE TABLE IF NOT EXISTS skills (
+            skill_key TEXT PRIMARY KEY, enabled INTEGER NOT NULL DEFAULT 1, updated_at TEXT NOT NULL
+        )
+    """)
+    conn.execute("""
+        CREATE VIRTUAL TABLE IF NOT EXISTS memories USING fts5(
+            fact, topic, source, created_at UNINDEXED
+        )
+    """)
+
+    if not conn.execute("SELECT id FROM profile WHERE id = 1").fetchone():
+        now = datetime.now(timezone.utc).isoformat()
+        conn.execute("INSERT INTO profile (id, content, updated_at) VALUES (1, ?, ?)", (DEFAULT_PROFILE, now))
+
+    if conn.execute("SELECT COUNT(*) as c FROM system_presets").fetchone()["c"] == 0:
+        now = datetime.now(timezone.utc).isoformat()
+        for preset in DEFAULT_PRESETS:
+            conn.execute(
+                "INSERT INTO system_presets (id, name, prompt, is_default, created_at) VALUES (?, ?, ?, 1, ?)",
+                (str(uuid.uuid4()), preset["name"], preset["prompt"], now),
+            )
+
+    defaults = {
+        "profile_enabled": "true", "default_model": DEFAULT_MODEL,
+        "search_enabled": "true", "memory_enabled": "true", "skills_enabled": "true",
+    }
+    for key, value in defaults.items():
+        if not conn.execute("SELECT key FROM settings WHERE key = ?", (key,)).fetchone():
+            conn.execute("INSERT INTO settings (key, value) VALUES (?, ?)", (key, value))
+
+    now = datetime.now(timezone.utc).isoformat()
+    for skill in BUILTIN_SKILLS:
+        if not conn.execute("SELECT skill_key FROM skills WHERE skill_key = ?", (skill["key"],)).fetchone():
+            conn.execute("INSERT INTO skills (skill_key, enabled, updated_at) VALUES (?, 1, ?)", (skill["key"], now))
+
+    existing_pin_hash = conn.execute("SELECT value FROM settings WHERE key = 'admin_pin_hash'").fetchone()
+    existing_pin_salt = conn.execute("SELECT value FROM settings WHERE key = 'admin_pin_salt'").fetchone()
+    if not existing_pin_hash or not existing_pin_salt:
+        from config import ALLOW_DEFAULT_PIN
+        configured_pin = os.getenv("JARVISCHAT_ADMIN_PIN", "").strip()
+        if re.fullmatch(r"\d{4}", configured_pin):
+            seed_pin, pin_source = configured_pin, "env"
+        elif ALLOW_DEFAULT_PIN:
+            seed_pin, pin_source = "1234", "default"
+        else:
+            raise RuntimeError(
+                "Admin PIN bootstrap blocked: set JARVISCHAT_ADMIN_PIN to a 4-digit PIN "
+                "or set JARVISCHAT_ALLOW_DEFAULT_PIN=true."
+            )
+        salt_hex, pin_hash_hex = hash_pin(seed_pin)
+        conn.execute("INSERT OR REPLACE INTO settings (key, value) VALUES (?, ?)", ("admin_pin_hash", pin_hash_hex))
+        conn.execute("INSERT OR REPLACE INTO settings (key, value) VALUES (?, ?)", ("admin_pin_salt", salt_hex))
+        if pin_source == "default":
+            log.warning("Admin PIN seeded from insecure default 1234 (override enabled).")
+        else:
+            log.info("Admin PIN hash seeded from configured environment PIN.")
+
+    conn.commit()
+    conn.close()

gpu.py

@@ -0,0 +1,31 @@
+"""
+JarvisChat - AMD GPU stats via rocm-smi.
+"""
+import json
+import logging
+import subprocess
+
+log = logging.getLogger("jarvischat")
+
+
+def get_gpu_stats() -> dict:
+    try:
+        result = subprocess.run(
+            ["rocm-smi", "--showuse", "--showmemuse", "--json"],
+            capture_output=True, text=True, timeout=5,
+        )
+        if result.returncode == 0:
+            data = json.loads(result.stdout)
+            gpu_info = data.get("card0", {})
+            gpu_use = gpu_info.get("GPU use (%)", 0)
+            vram_use = gpu_info.get("GPU Memory Allocated (VRAM%)", 0)
+            if isinstance(gpu_use, str):
+                gpu_use = int(gpu_use.replace("%", "").strip() or 0)
+            if isinstance(vram_use, str):
+                vram_use = int(vram_use.replace("%", "").strip() or 0)
+            return {"gpu_percent": gpu_use, "vram_percent": vram_use, "available": True}
+    except (subprocess.TimeoutExpired, FileNotFoundError, json.JSONDecodeError):
+        pass
+    except Exception as e:
+        log.warning(f"GPU stats error: {e}")
+    return {"gpu_percent": 0, "vram_percent": 0, "available": False}

memory.py

@@ -0,0 +1,139 @@
+"""
+JarvisChat - FTS5 memory system.
+CRUD, search, remember/forget command processing, topic detection.
+"""
+import logging
+import re
+from datetime import datetime, timezone
+from typing import Optional
+
+from db import get_db
+from config import MAX_MEMORY_FACT_CHARS
+
+log = logging.getLogger("jarvischat")
+
+REMEMBER_PATTERNS = [
+    (r"remember that (.+)", "explicit"),
+    (r"please remember (.+)", "explicit"),
+    (r"don'?t forget (.+)", "explicit"),
+    (r"note that (.+)", "explicit"),
+    (r"keep in mind (?:that )?(.+)", "explicit"),
+]
+
+FORGET_PATTERNS = [
+    r"forget (?:that )?(.+)",
+    r"don'?t remember (.+)",
+    r"remove (?:the )?memory (?:about |that )?(.+)",
+]
+
+
+def detect_topic(fact: str) -> str:
+    fact_lower = fact.lower()
+    if any(w in fact_lower for w in ["prefer", "like", "hate", "always", "never", "favorite"]):
+        return "preference"
+    elif any(w in fact_lower for w in ["working on", "building", "project", "developing"]):
+        return "project"
+    elif any(w in fact_lower for w in ["run", "install", "server", "ip", "port", "service", "docker", "systemd"]):
+        return "infrastructure"
+    elif any(w in fact_lower for w in ["my name", "i am", "i'm a", "i live", "my wife", "my partner"]):
+        return "personal"
+    return "general"
+
+
+def add_memory(fact: str, topic: str = "general", source: str = "explicit") -> Optional[int]:
+    db = get_db()
+    now = datetime.now(timezone.utc).isoformat()
+    cur = db.execute(
+        "INSERT INTO memories (fact, topic, source, created_at) VALUES (?, ?, ?, ?)",
+        (fact, topic, source, now),
+    )
+    db.commit()
+    rowid = cur.lastrowid
+    db.close()
+    log.info(f"Memory added [{topic}]: {fact[:50]}...")
+    return rowid
+
+
+def search_memories(query: str, limit: int = 5) -> list:
+    if not query.strip():
+        return []
+    db = get_db()
+    words = re.findall(r"[A-Za-z0-9_]+", query)
+    if not words:
+        db.close()
+        return []
+    safe_query = " OR ".join(word + "*" for word in words[:10])
+    try:
+        rows = db.execute(
+            "SELECT rowid, fact, topic, source, created_at, bm25(memories) AS rank "
+            "FROM memories WHERE memories MATCH ? ORDER BY rank LIMIT ?",
+            (safe_query, limit),
+        ).fetchall()
+        results = [dict(row) for row in rows]
+        log.debug(f"Memory search '{query}' returned {len(results)} results")
+    except Exception as e:
+        log.warning(f"Memory search error: {e}")
+        results = []
+    db.close()
+    return results
+
+
+def get_all_memories(topic: Optional[str] = None) -> list:
+    db = get_db()
+    if topic:
+        rows = db.execute(
+            "SELECT rowid, * FROM memories WHERE topic = ? ORDER BY created_at DESC", (topic,)
+        ).fetchall()
+    else:
+        rows = db.execute("SELECT rowid, * FROM memories ORDER BY created_at DESC").fetchall()
+    db.close()
+    return [dict(row) for row in rows]
+
+
+def delete_memory(rowid: int) -> bool:
+    db = get_db()
+    cur = db.execute("DELETE FROM memories WHERE rowid = ?", (rowid,))
+    db.commit()
+    deleted = cur.rowcount > 0
+    db.close()
+    if deleted:
+        log.info(f"Memory deleted: rowid={rowid}")
+    return deleted
+
+
+def update_memory(rowid: int, fact: str) -> bool:
+    db = get_db()
+    cur = db.execute("UPDATE memories SET fact = ? WHERE rowid = ?", (fact, rowid))
+    db.commit()
+    updated = cur.rowcount > 0
+    db.close()
+    return updated
+
+
+def get_memory_count() -> int:
+    db = get_db()
+    count = db.execute("SELECT COUNT(*) as c FROM memories").fetchone()["c"]
+    db.close()
+    return count
+
+
+def process_remember_command(user_message: str) -> Optional[str]:
+    for pattern, source in REMEMBER_PATTERNS:
+        match = re.search(pattern, user_message, re.IGNORECASE)
+        if match:
+            fact = match.group(1).strip().rstrip(".")
+            topic = detect_topic(fact)
+            add_memory(fact, topic=topic, source=source)
+            return f"✓ Remembered [{topic}]: {fact}"
+    for pattern in FORGET_PATTERNS:
+        match = re.search(pattern, user_message, re.IGNORECASE)
+        if match:
+            search_term = match.group(1).strip().rstrip(".")
+            memories = search_memories(search_term, limit=3)
+            if memories:
+                for m in memories:
+                    delete_memory(m["rowid"])
+                return f"✓ Forgot {len(memories)} memory/memories about: {search_term}"
+            else:
+                return f"✗ No memories found about: {search_term}"
+    return None

rag.py

@@ -0,0 +1,80 @@
+"""
+JarvisChat - RAG pipeline: Qdrant vector search + system prompt assembly.
+"""
+import logging
+
+import httpx
+
+from db import get_db, get_setting, list_skills_with_state, format_active_skills_prompt
+from memory import search_memories
+from config import MAX_SKILL_PROMPT_CHARS
+
+log = logging.getLogger("jarvischat")
+
+QDRANT_URL = "http://192.168.50.108:6333"
+EMBED_URL = "http://192.168.50.108:11434"
+EMBED_MODEL = "mxbai-embed-large"
+RAG_COLLECTION = "jarvis_rag"
+RAG_SCORE_THRESHOLD = 0.25
+
+
+async def query_rag(query: str, limit: int = 3) -> list:
+    try:
+        async with httpx.AsyncClient() as client:
+            embed_resp = await client.post(
+                f"{EMBED_URL}/api/embeddings",
+                json={"model": EMBED_MODEL, "prompt": query},
+                timeout=10.0,
+            )
+            if embed_resp.status_code != 200:
+                return []
+            vector = embed_resp.json()["embedding"]
+            search_resp = await client.post(
+                f"{QDRANT_URL}/collections/{RAG_COLLECTION}/points/search",
+                json={"vector": vector, "limit": limit, "with_payload": True},
+                timeout=10.0,
+            )
+            if search_resp.status_code != 200:
+                return []
+            return search_resp.json().get("result", [])
+    except Exception as e:
+        log.warning(f"RAG query error: {e}")
+        return []
+
+
+async def build_system_prompt(db, extra_prompt: str = "", user_message: str = "") -> str:
+    parts = []
+    settings = {row["key"]: row["value"] for row in db.execute("SELECT key, value FROM settings").fetchall()}
+
+    if settings.get("profile_enabled", "true") == "true":
+        profile = db.execute("SELECT content FROM profile WHERE id = 1").fetchone()
+        if profile and profile["content"].strip():
+            parts.append(profile["content"].strip())
+
+    if settings.get("memory_enabled", "true") == "true" and user_message:
+        memories = search_memories(user_message, limit=5)
+        if memories:
+            memory_lines = [f"- {m['fact']}" for m in memories]
+            parts.append("## Relevant Context from Memory\n" + "\n".join(memory_lines))
+            log.debug(f"Injected {len(memories)} memories into context")
+
+    if user_message:
+        try:
+            rag_results = await query_rag(user_message)
+            if rag_results:
+                rag_lines = [r["payload"]["text"] for r in rag_results if r["score"] > RAG_SCORE_THRESHOLD]
+                if rag_lines:
+                    parts.append("## Retrieved Context\n" + "\n\n---\n\n".join(rag_lines))
+                    log.warning(f"RAG injected {len(rag_lines)} chunks into context")
+        except Exception as e:
+            log.warning(f"RAG injection error: {e}")
+
+    if settings.get("skills_enabled", "true") == "true":
+        active_skills = [s for s in list_skills_with_state(db) if s["enabled"]]
+        if active_skills:
+            parts.append(format_active_skills_prompt(active_skills))
+
+    if extra_prompt and extra_prompt.strip():
+        parts.append(extra_prompt.strip())
+
+    return "\n\n---\n\n".join(parts) if parts else ""

routers/chat.py

@@ -0,0 +1,203 @@
+"""JarvisChat routers - /api/chat streaming endpoint."""
+import json
+import logging
+import uuid
+from datetime import datetime, timezone
+
+import httpx
+from fastapi import APIRouter, HTTPException, Request
+from fastapi.responses import StreamingResponse
+
+from config import DEFAULT_MODEL, LLAMA_SERVER_BASE
+from db import get_db
+from memory import process_remember_command
+from rag import build_system_prompt
+from search import (calculate_perplexity, is_uncertain, is_refusal,
+                    clean_hedging, format_search_results, format_direct_answer,
+                    extract_search_query, query_searxng)
+from security import read_json_body, log_incident, BODY_LIMIT_CHAT_BYTES
+from config import MAX_CHAT_MESSAGE_CHARS
+
+log = logging.getLogger("jarvischat")
+router = APIRouter()
+
+
+def parse_llama_stream_chunk(line: str) -> tuple:
+    if line.startswith("data: "):
+        line = line[6:]
+    if line.strip() == "[DONE]":
+        return None, True, {}
+    try:
+        chunk = json.loads(line)
+        choices = chunk.get("choices", [])
+        if choices:
+            delta = choices[0].get("delta", {})
+            token = delta.get("content")
+            finish = choices[0].get("finish_reason")
+            stats = {}
+            if finish == "stop":
+                usage = chunk.get("usage", {})
+                stats["tokens_per_sec"] = usage.get("tokens_per_second", 0.0)
+            return token, finish == "stop", stats
+        if "message" in chunk and "content" in chunk["message"]:
+            token = chunk["message"]["content"]
+            done = chunk.get("done", False)
+            stats = {}
+            if done:
+                eval_count = chunk.get("eval_count", 0)
+                eval_duration = chunk.get("eval_duration", 0)
+                stats["tokens_per_sec"] = (eval_count / (eval_duration / 1e9)) if eval_duration > 0 else 0
+            return token, done, stats
+    except json.JSONDecodeError:
+        pass
+    return None, False, {}
+
+
+@router.post("/api/chat")
+async def chat(request: Request):
+    body = await read_json_body(request, BODY_LIMIT_CHAT_BYTES)
+    conv_id = body.get("conversation_id")
+    user_message = body.get("message", "").strip()
+    if len(user_message) > MAX_CHAT_MESSAGE_CHARS:
+        raise HTTPException(status_code=413, detail="Chat message is too long")
+    model = body.get("model", DEFAULT_MODEL)
+    preset_prompt = body.get("system_prompt", "")
+
+    if not user_message:
+        raise HTTPException(status_code=400, detail="Empty message")
+
+    db = get_db()
+    now = datetime.now(timezone.utc).isoformat()
+    settings = {row["key"]: row["value"] for row in db.execute("SELECT key, value FROM settings").fetchall()}
+    search_enabled = settings.get("search_enabled", "true") == "true"
+
+    remember_response = process_remember_command(user_message)
+
+    if not conv_id:
+        conv_id = str(uuid.uuid4())
+        title = user_message[:80] + ("..." if len(user_message) > 80 else "")
+        db.execute("INSERT INTO conversations (id, title, model, created_at, updated_at) VALUES (?, ?, ?, ?, ?)",
+                   (conv_id, title, model, now, now))
+    else:
+        db.execute("UPDATE conversations SET updated_at = ? WHERE id = ?", (now, conv_id))
+
+    db.execute("INSERT INTO messages (conversation_id, role, content, created_at) VALUES (?, ?, ?, ?)",
+               (conv_id, "user", user_message, now))
+    db.commit()
+
+    history_rows = db.execute(
+        "SELECT role, content FROM messages WHERE conversation_id = ? ORDER BY id ASC", (conv_id,)
+    ).fetchall()
+    system_prompt = await build_system_prompt(db, preset_prompt, user_message)
+    db.close()
+
+    messages = []
+    if system_prompt:
+        messages.append({"role": "system", "content": system_prompt})
+    for row in history_rows:
+        messages.append({"role": row["role"], "content": row["content"]})
+
+    ollama_payload = {"model": model, "messages": messages, "stream": True}
+
+    async def stream_response():
+        full_response = []
+        all_logprobs = []
+        tokens_per_sec = 0.0
+
+        if remember_response:
+            yield f"data: {json.dumps({'token': remember_response + chr(10) + chr(10), 'conversation_id': conv_id})}\n\n"
+
+        async with httpx.AsyncClient() as client:
+            try:
+                async with client.stream(
+                    "POST", f"{LLAMA_SERVER_BASE}/v1/chat/completions",
+                    json=ollama_payload,
+                    timeout=httpx.Timeout(300.0, connect=10.0),
+                ) as resp:
+                    async for line in resp.aiter_lines():
+                        if line.strip():
+                            token, done, stats = parse_llama_stream_chunk(line)
+                            if token:
+                                full_response.append(token)
+                                yield f"data: {json.dumps({'token': token, 'conversation_id': conv_id})}\n\n"
+                            if done:
+                                tokens_per_sec = stats.get("tokens_per_sec", 0.0)
+
+                assistant_msg = "".join(full_response)
+                perplexity = calculate_perplexity(all_logprobs) if all_logprobs else 0.0
+                should_search = is_uncertain(all_logprobs) or is_refusal(assistant_msg)
+
+                if search_enabled and should_search:
+                    yield f"data: {json.dumps({'searching': True, 'conversation_id': conv_id})}\n\n"
+                    search_query = extract_search_query(user_message)
+                    search_results = await query_searxng(search_query)
+
+                    if search_results:
+                        search_context = format_search_results(search_results)
+                        augmented_messages = []
+                        if system_prompt:
+                            augmented_messages.append({"role": "system", "content": system_prompt + "\n\n" + search_context})
+                        else:
+                            augmented_messages.append({"role": "system", "content": search_context})
+                        for row in history_rows[:-1]:
+                            augmented_messages.append({"role": row["role"], "content": row["content"]})
+                        augmented_messages.append({"role": "user", "content": user_message})
+
+                        yield f"data: {json.dumps({'search_results': len(search_results), 'conversation_id': conv_id})}\n\n"
+
+                        augmented_response = []
+                        async with client.stream(
+                            "POST", f"{LLAMA_SERVER_BASE}/v1/chat/completions",
+                            json={"model": model, "messages": augmented_messages, "stream": True},
+                            timeout=httpx.Timeout(300.0, connect=10.0),
+                        ) as resp2:
+                            async for line in resp2.aiter_lines():
+                                if line.strip():
+                                    token2, done2, _ = parse_llama_stream_chunk(line)
+                                    if token2:
+                                        augmented_response.append(token2)
+                                    if done2:
+                                        break
+
+                        raw_response = "".join(augmented_response) or assistant_msg
+                        cleaned_response = clean_hedging(raw_response)
+                        if is_refusal(cleaned_response) or len(cleaned_response) < 20:
+                            cleaned_response = format_direct_answer(user_message, search_results)
+
+                        yield f"data: {json.dumps({'token': cleaned_response, 'conversation_id': conv_id, 'augmented': True})}\n\n"
+
+                        saved_msg = cleaned_response + "\n\n---\n*🔍 Enhanced with web search results*"
+                        if remember_response:
+                            saved_msg = remember_response + "\n\n" + saved_msg
+
+                        db2 = get_db()
+                        db2.execute("INSERT INTO messages (conversation_id, role, content, created_at) VALUES (?, ?, ?, ?)",
+                                    (conv_id, "assistant", saved_msg, datetime.now(timezone.utc).isoformat()))
+                        db2.commit()
+                        db2.close()
+
+                        yield f"data: {json.dumps({'done': True, 'conversation_id': conv_id, 'searched': True, 'perplexity': round(perplexity, 2), 'tokens_per_sec': round(tokens_per_sec, 1)})}\n\n"
+                        return
+
+                saved_msg = assistant_msg
+                if remember_response:
+                    saved_msg = remember_response + "\n\n" + saved_msg
+
+                db2 = get_db()
+                db2.execute("INSERT INTO messages (conversation_id, role, content, created_at) VALUES (?, ?, ?, ?)",
+                            (conv_id, "assistant", saved_msg, datetime.now(timezone.utc).isoformat()))
+                db2.commit()
+                db2.close()
+
+                yield f"data: {json.dumps({'done': True, 'conversation_id': conv_id, 'perplexity': round(perplexity, 2), 'tokens_per_sec': round(tokens_per_sec, 1)})}\n\n"
+
+            except httpx.RemoteProtocolError:
+                pass
+            except httpx.ConnectError:
+                yield f"data: {json.dumps({'error': 'Cannot connect to Ollama. Is it running?'})}\n\n"
+            except Exception as e:
+                incident_key = log_incident("chat_stream", message="Ollama stream failure during chat response",
+                                            request=request, exc=e)
+                yield f"data: {json.dumps({'error': 'Chat response generation failed before completion. Use the incident key for support lookup.', 'error_key': incident_key})}\n\n"
+
+    return StreamingResponse(stream_response(), media_type="text/event-stream")

routers/completions.py

@@ -0,0 +1,267 @@
+"""
+JarvisChat - /v1/chat/completions router.
+OpenAI-compatible endpoint for IDE integration (Continue.dev, etc.).
+Runs all requests through the full jC pipeline: profile + RAG + memory injection.
+FIM (fill-in-the-middle) requests are proxied directly — not persisted.
+Chat-style requests are persisted to conversation history.
+Auth: static Bearer token via COMPLETIONS_API_KEY in config.
+"""
+import json
+import logging
+import uuid
+from datetime import datetime, timezone
+
+import httpx
+from fastapi import APIRouter, HTTPException, Request
+from fastapi.responses import StreamingResponse, JSONResponse
+
+from config import DEFAULT_MODEL, LLAMA_SERVER_BASE, COMPLETIONS_API_KEY
+from db import get_db
+from rag import build_system_prompt
+from routers.chat import parse_llama_stream_chunk
+
+log = logging.getLogger("jarvischat")
+router = APIRouter()
+
+
+def _check_api_key(request: Request):
+    auth = request.headers.get("Authorization", "")
+    if not auth.startswith("Bearer "):
+        raise HTTPException(status_code=401, detail="Missing Bearer token")
+    token = auth[7:].strip()
+    if token != COMPLETIONS_API_KEY:
+        raise HTTPException(status_code=401, detail="Invalid API key")
+
+
+def _is_fim_request(body: dict) -> bool:
+    """
+    FIM (fill-in-the-middle) requests use a 'prompt' + optional 'suffix' structure
+    rather than a 'messages' array. Continue.dev sends these for inline autocomplete.
+    We proxy them directly without pipeline injection or persistence.
+    """
+    return "prompt" in body and "messages" not in body
+
+
+def _build_openai_chunk(token: str, model: str, conv_id: str) -> str:
+    chunk = {
+        "id": f"chatcmpl-{conv_id}",
+        "object": "chat.completion.chunk",
+        "model": model,
+        "choices": [{
+            "index": 0,
+            "delta": {"content": token},
+            "finish_reason": None,
+        }],
+    }
+    return f"data: {json.dumps(chunk)}\n\n"
+
+
+def _build_openai_stop_chunk(model: str, conv_id: str) -> str:
+    chunk = {
+        "id": f"chatcmpl-{conv_id}",
+        "object": "chat.completion.chunk",
+        "model": model,
+        "choices": [{
+            "index": 0,
+            "delta": {},
+            "finish_reason": "stop",
+        }],
+    }
+    return f"data: {json.dumps(chunk)}\n\n"
+
+
+def _build_openai_response(content: str, model: str, conv_id: str) -> dict:
+    """Non-streaming response envelope."""
+    return {
+        "id": f"chatcmpl-{conv_id}",
+        "object": "chat.completion",
+        "model": model,
+        "choices": [{
+            "index": 0,
+            "message": {"role": "assistant", "content": content},
+            "finish_reason": "stop",
+        }],
+        "usage": {"prompt_tokens": 0, "completion_tokens": 0, "total_tokens": 0},
+    }
+
+
+@router.post("/v1/chat/completions")
+async def chat_completions(request: Request):
+    _check_api_key(request)
+
+    try:
+        body = await request.json()
+    except Exception:
+        raise HTTPException(status_code=400, detail="Invalid JSON body")
+
+    # --- FIM passthrough ---
+    if _is_fim_request(body):
+        return await _fim_passthrough(body)
+
+    # --- Chat completion ---
+    messages = body.get("messages", [])
+    if not messages:
+        raise HTTPException(status_code=400, detail="No messages provided")
+
+    model = body.get("model", DEFAULT_MODEL)
+    stream = body.get("stream", True)
+
+    # Extract the latest user message for RAG + conversation title
+    user_message = ""
+    for msg in reversed(messages):
+        if msg.get("role") == "user":
+            user_message = msg.get("content", "").strip()
+            break
+
+    if not user_message:
+        raise HTTPException(status_code=400, detail="No user message found")
+
+    # --- Persist conversation ---
+    db = get_db()
+    now = datetime.now(timezone.utc).isoformat()
+    conv_id = str(uuid.uuid4())
+    title = f"[IDE] {user_message[:72]}{'...' if len(user_message) > 72 else ''}"
+    db.execute(
+        "INSERT INTO conversations (id, title, model, created_at, updated_at) VALUES (?, ?, ?, ?, ?)",
+        (conv_id, title, model, now, now),
+    )
+    for msg in messages:
+        role = msg.get("role")
+        content = msg.get("content", "")
+        if role in ("user", "assistant"):
+            db.execute(
+                "INSERT INTO messages (conversation_id, role, content, created_at) VALUES (?, ?, ?, ?)",
+                (conv_id, role, content, now),
+            )
+    db.commit()
+
+    # --- Build system prompt through full jC pipeline ---
+    system_prompt = await build_system_prompt(db, "", user_message)
+    db.close()
+
+    # Assemble messages for upstream: inject jC system prompt, preserve history
+    upstream_messages = []
+    if system_prompt:
+        upstream_messages.append({"role": "system", "content": system_prompt})
+
+    # Strip any system messages from the incoming payload — jC owns the system prompt
+    for msg in messages:
+        if msg.get("role") != "system":
+            upstream_messages.append(msg)
+
+    upstream_payload = {
+        "model": model,
+        "messages": upstream_messages,
+        "stream": True,  # always stream from upstream; we buffer if client wants non-stream
+    }
+
+    if stream:
+        return StreamingResponse(
+            _stream_chat(upstream_payload, model, conv_id, request),
+            media_type="text/event-stream",
+        )
+    else:
+        return await _blocking_chat(upstream_payload, model, conv_id, request)
+
+
+async def _stream_chat(payload: dict, model: str, conv_id: str, request: Request):
+    """Stream tokens to client in OpenAI SSE format, persist assistant response."""
+    full_response = []
+
+    async with httpx.AsyncClient() as client:
+        try:
+            async with client.stream(
+                "POST", f"{LLAMA_SERVER_BASE}/v1/chat/completions",
+                json=payload,
+                timeout=httpx.Timeout(300.0, connect=10.0),
+            ) as resp:
+                async for line in resp.aiter_lines():
+                    if not line.strip():
+                        continue
+                    token, done, _ = parse_llama_stream_chunk(line)
+                    if token:
+                        full_response.append(token)
+                        yield _build_openai_chunk(token, model, conv_id)
+                    if done:
+                        break
+
+            yield _build_openai_stop_chunk(model, conv_id)
+            yield "data: [DONE]\n\n"
+
+            # Persist assistant response
+            assistant_msg = "".join(full_response)
+            if assistant_msg:
+                db = get_db()
+                db.execute(
+                    "INSERT INTO messages (conversation_id, role, content, created_at) VALUES (?, ?, ?, ?)",
+                    (conv_id, "assistant", assistant_msg, datetime.now(timezone.utc).isoformat()),
+                )
+                db.commit()
+                db.close()
+
+        except httpx.ConnectError:
+            err = {"error": {"message": "Cannot connect to inference server", "type": "connection_error"}}
+            yield f"data: {json.dumps(err)}\n\n"
+        except Exception as e:
+            log.error(f"completions stream error: {e}")
+            err = {"error": {"message": "Stream failed", "type": "server_error"}}
+            yield f"data: {json.dumps(err)}\n\n"
+
+
+async def _blocking_chat(payload: dict, model: str, conv_id: str, request: Request) -> JSONResponse:
+    """Accumulate full response, return as standard OpenAI JSON object."""
+    full_response = []
+
+    async with httpx.AsyncClient() as client:
+        try:
+            async with client.stream(
+                "POST", f"{LLAMA_SERVER_BASE}/v1/chat/completions",
+                json=payload,
+                timeout=httpx.Timeout(300.0, connect=10.0),
+            ) as resp:
+                async for line in resp.aiter_lines():
+                    if not line.strip():
+                        continue
+                    token, done, _ = parse_llama_stream_chunk(line)
+                    if token:
+                        full_response.append(token)
+                    if done:
+                        break
+        except httpx.ConnectError:
+            raise HTTPException(status_code=503, detail="Cannot connect to inference server")
+        except Exception as e:
+            log.error(f"completions blocking error: {e}")
+            raise HTTPException(status_code=500, detail="Inference request failed")
+
+    assistant_msg = "".join(full_response)
+
+    if assistant_msg:
+        db = get_db()
+        db.execute(
+            "INSERT INTO messages (conversation_id, role, content, created_at) VALUES (?, ?, ?, ?)",
+            (conv_id, "assistant", assistant_msg, datetime.now(timezone.utc).isoformat()),
+        )
+        db.commit()
+        db.close()
+
+    return JSONResponse(content=_build_openai_response(assistant_msg, model, conv_id))
+
+
+async def _fim_passthrough(body: dict) -> JSONResponse:
+    """
+    Proxy FIM requests directly to llama-server without pipeline injection.
+    Not persisted — autocomplete noise has no RAG value.
+    """
+    async with httpx.AsyncClient() as client:
+        try:
+            resp = await client.post(
+                f"{LLAMA_SERVER_BASE}/v1/completions",
+                json=body,
+                timeout=httpx.Timeout(30.0, connect=5.0),
+            )
+            return JSONResponse(content=resp.json(), status_code=resp.status_code)
+        except httpx.ConnectError:
+            raise HTTPException(status_code=503, detail="Cannot connect to inference server")
+        except Exception as e:
+            log.error(f"FIM passthrough error: {e}")
+            raise HTTPException(status_code=500, detail="FIM request failed")

routers/conversations.py

@@ -0,0 +1,83 @@
+"""JarvisChat routers - Conversation CRUD."""
+import logging
+import uuid
+from datetime import datetime, timezone
+from fastapi import APIRouter, HTTPException, Request
+from db import get_db
+from security import read_json_body, BODY_LIMIT_DEFAULT_BYTES
+from config import DEFAULT_MODEL, MAX_CONVERSATION_TITLE_CHARS
+
+log = logging.getLogger("jarvischat")
+router = APIRouter()
+
+
+@router.get("/api/conversations")
+async def list_conversations():
+    db = get_db()
+    rows = db.execute("SELECT * FROM conversations ORDER BY updated_at DESC").fetchall()
+    db.close()
+    return [dict(r) for r in rows]
+
+
+@router.post("/api/conversations")
+async def create_conversation(request: Request):
+    body = await read_json_body(request, BODY_LIMIT_DEFAULT_BYTES)
+    conv_id = str(uuid.uuid4())
+    now = datetime.now(timezone.utc).isoformat()
+    model = body.get("model", DEFAULT_MODEL)
+    title = str(body.get("title", "New Chat"))[:MAX_CONVERSATION_TITLE_CHARS]
+    db = get_db()
+    db.execute("INSERT INTO conversations (id, title, model, created_at, updated_at) VALUES (?, ?, ?, ?, ?)",
+               (conv_id, title, model, now, now))
+    db.commit()
+    db.close()
+    return {"id": conv_id, "title": title, "model": model, "created_at": now, "updated_at": now}
+
+
+@router.get("/api/conversations/{conv_id}")
+async def get_conversation(conv_id: str):
+    db = get_db()
+    conv = db.execute("SELECT * FROM conversations WHERE id = ?", (conv_id,)).fetchone()
+    if not conv:
+        db.close()
+        raise HTTPException(status_code=404, detail="Conversation not found")
+    messages = db.execute("SELECT * FROM messages WHERE conversation_id = ? ORDER BY id ASC", (conv_id,)).fetchall()
+    db.close()
+    return {"conversation": dict(conv), "messages": [dict(m) for m in messages]}
+
+
+@router.put("/api/conversations/{conv_id}")
+async def update_conversation(conv_id: str, request: Request):
+    body = await read_json_body(request, BODY_LIMIT_DEFAULT_BYTES)
+    db = get_db()
+    now = datetime.now(timezone.utc).isoformat()
+    if "title" in body:
+        db.execute("UPDATE conversations SET title = ?, updated_at = ? WHERE id = ?",
+                   (str(body["title"])[:MAX_CONVERSATION_TITLE_CHARS], now, conv_id))
+    if "model" in body:
+        db.execute("UPDATE conversations SET model = ?, updated_at = ? WHERE id = ?",
+                   (body["model"], now, conv_id))
+    db.commit()
+    db.close()
+    return {"status": "ok"}
+
+
+@router.delete("/api/conversations/{conv_id}")
+async def delete_conversation(conv_id: str):
+    db = get_db()
+    db.execute("DELETE FROM messages WHERE conversation_id = ?", (conv_id,))
+    db.execute("DELETE FROM conversations WHERE id = ?", (conv_id,))
+    db.commit()
+    db.close()
+    return {"status": "ok"}
+
+
+@router.delete("/api/conversations")
+async def delete_all_conversations():
+    db = get_db()
+    db.execute("DELETE FROM messages")
+    db.execute("DELETE FROM conversations")
+    db.commit()
+    db.close()
+    log.info("Deleted all conversations")
+    return {"status": "ok"}

routers/memories.py

@@ -0,0 +1,63 @@
+"""JarvisChat routers - Memory CRUD API."""
+from fastapi import APIRouter, HTTPException, Request
+from typing import Optional
+
+from db import get_db
+from memory import add_memory, delete_memory, update_memory, get_all_memories, search_memories
+from security import read_json_body, BODY_LIMIT_DEFAULT_BYTES
+from config import MAX_MEMORY_FACT_CHARS
+
+router = APIRouter()
+
+
+@router.get("/api/memories")
+async def list_memories(topic: Optional[str] = None):
+    memories = get_all_memories(topic)
+    return {"memories": memories, "count": len(memories)}
+
+
+@router.post("/api/memories")
+async def create_memory(request: Request):
+    body = await read_json_body(request, BODY_LIMIT_DEFAULT_BYTES)
+    fact = str(body.get("fact", "")).strip()
+    if not fact:
+        raise HTTPException(status_code=400, detail="Memory fact is required")
+    if len(fact) > MAX_MEMORY_FACT_CHARS:
+        raise HTTPException(status_code=413, detail="Memory fact is too long")
+    rowid = add_memory(fact=fact, topic=body.get("topic", "general"), source=body.get("source", "manual"))
+    return {"rowid": rowid, "status": "ok"}
+
+
+@router.delete("/api/memories/{rowid}")
+async def remove_memory(rowid: int):
+    if not delete_memory(rowid):
+        raise HTTPException(status_code=404, detail="Memory not found")
+    return {"status": "ok"}
+
+
+@router.put("/api/memories/{rowid}")
+async def edit_memory(rowid: int, request: Request):
+    body = await read_json_body(request, BODY_LIMIT_DEFAULT_BYTES)
+    fact = str(body.get("fact", "")).strip()
+    if not fact:
+        raise HTTPException(status_code=400, detail="Memory fact is required")
+    if len(fact) > MAX_MEMORY_FACT_CHARS:
+        raise HTTPException(status_code=413, detail="Memory fact is too long")
+    if not update_memory(rowid, fact):
+        raise HTTPException(status_code=404, detail="Memory not found")
+    return {"status": "ok"}
+
+
+@router.get("/api/memories/search")
+async def search_memories_api(q: str, limit: int = 10):
+    results = search_memories(q, limit=limit)
+    return {"results": results, "count": len(results)}
+
+
+@router.get("/api/memories/stats")
+async def memory_stats():
+    db = get_db()
+    total = db.execute("SELECT COUNT(*) as c FROM memories").fetchone()["c"]
+    topics = db.execute("SELECT topic, COUNT(*) as c FROM memories GROUP BY topic ORDER BY c DESC").fetchall()
+    db.close()
+    return {"total": total, "by_topic": {row["topic"]: row["c"] for row in topics}}

routers/models.py

@@ -0,0 +1,78 @@
+"""
+JarvisChat routers - Model listing, system stats.
+"""
+import logging
+from typing import Optional
+
+import httpx
+import psutil
+from fastapi import APIRouter, HTTPException, Request
+
+from config import OLLAMA_BASE
+from gpu import get_gpu_stats
+from security import read_json_body, BODY_LIMIT_DEFAULT_BYTES
+
+log = logging.getLogger("jarvischat")
+router = APIRouter()
+
+
+@router.get("/api/models")
+async def list_models():
+    async with httpx.AsyncClient() as client:
+        try:
+            resp = await client.get(f"{OLLAMA_BASE}/v1/models", timeout=10)
+            data = resp.json()
+            models = [{"name": m["id"], "model": m["id"]} for m in data.get("data", [])]
+            return {"models": models}
+        except httpx.ConnectError:
+            raise HTTPException(status_code=502, detail="Cannot connect to llama-server.")
+
+
+@router.get("/api/ps")
+async def running_models():
+    async with httpx.AsyncClient() as client:
+        try:
+            resp = await client.get(f"{OLLAMA_BASE}/api/ps", timeout=10)
+            return resp.json()
+        except httpx.ConnectError:
+            raise HTTPException(status_code=502, detail="Cannot connect to Ollama.")
+
+
+@router.post("/api/show")
+async def show_model(request: Request):
+    from security import BODY_LIMIT_DEFAULT_BYTES
+    body = await read_json_body(request, BODY_LIMIT_DEFAULT_BYTES)
+    async with httpx.AsyncClient() as client:
+        try:
+            resp = await client.post(f"{OLLAMA_BASE}/api/show", json=body, timeout=10)
+            return resp.json()
+        except httpx.ConnectError:
+            raise HTTPException(status_code=502, detail="Cannot connect to Ollama.")
+
+
+@router.get("/api/stats")
+async def system_stats():
+    cpu_percent = psutil.cpu_percent(interval=0.1)
+    memory = psutil.virtual_memory()
+    gpu = get_gpu_stats()
+    return {
+        "cpu_percent": round(cpu_percent, 1),
+        "memory_percent": round(memory.percent, 1),
+        "memory_used_gb": round(memory.used / (1024**3), 1),
+        "memory_total_gb": round(memory.total / (1024**3), 1),
+        "gpu_percent": gpu["gpu_percent"],
+        "vram_percent": gpu["vram_percent"],
+        "gpu_available": gpu["available"],
+    }
+
+
+@router.get("/api/search/status")
+async def search_status():
+    from config import SEARXNG_BASE
+    async with httpx.AsyncClient() as client:
+        try:
+            resp = await client.get(f"{SEARXNG_BASE}/search",
+                                    params={"q": "test", "format": "json"}, timeout=5)
+            return {"available": resp.status_code == 200}
+        except Exception:
+            return {"available": False}

routers/presets.py

@@ -0,0 +1,61 @@
+"""JarvisChat routers - System prompt presets."""
+import uuid
+from datetime import datetime, timezone
+from fastapi import APIRouter, HTTPException, Request
+from db import get_db
+from security import read_json_body, BODY_LIMIT_DEFAULT_BYTES
+from config import MAX_PRESET_NAME_CHARS, MAX_PRESET_PROMPT_CHARS
+
+router = APIRouter()
+
+
+@router.get("/api/presets")
+async def list_presets():
+    db = get_db()
+    rows = db.execute("SELECT * FROM system_presets ORDER BY is_default DESC, name ASC").fetchall()
+    db.close()
+    return [dict(r) for r in rows]
+
+
+@router.post("/api/presets")
+async def create_preset(request: Request):
+    body = await read_json_body(request, BODY_LIMIT_DEFAULT_BYTES)
+    name = str(body.get("name", "")).strip()
+    prompt = str(body.get("prompt", "")).strip()
+    if not name or not prompt:
+        raise HTTPException(status_code=400, detail="Preset name and prompt are required")
+    if len(name) > MAX_PRESET_NAME_CHARS or len(prompt) > MAX_PRESET_PROMPT_CHARS:
+        raise HTTPException(status_code=413, detail="Preset fields are too long")
+    preset_id = str(uuid.uuid4())
+    now = datetime.now(timezone.utc).isoformat()
+    db = get_db()
+    db.execute("INSERT INTO system_presets (id, name, prompt, is_default, created_at) VALUES (?, ?, ?, 0, ?)",
+               (preset_id, name, prompt, now))
+    db.commit()
+    db.close()
+    return {"id": preset_id, "name": name, "prompt": prompt}
+
+
+@router.put("/api/presets/{preset_id}")
+async def update_preset(preset_id: str, request: Request):
+    body = await read_json_body(request, BODY_LIMIT_DEFAULT_BYTES)
+    name = str(body.get("name", "")).strip()
+    prompt = str(body.get("prompt", "")).strip()
+    if not name or not prompt:
+        raise HTTPException(status_code=400, detail="Preset name and prompt are required")
+    if len(name) > MAX_PRESET_NAME_CHARS or len(prompt) > MAX_PRESET_PROMPT_CHARS:
+        raise HTTPException(status_code=413, detail="Preset fields are too long")
+    db = get_db()
+    db.execute("UPDATE system_presets SET name = ?, prompt = ? WHERE id = ?", (name, prompt, preset_id))
+    db.commit()
+    db.close()
+    return {"status": "ok"}
+
+
+@router.delete("/api/presets/{preset_id}")
+async def delete_preset(preset_id: str):
+    db = get_db()
+    db.execute("DELETE FROM system_presets WHERE id = ? AND is_default = 0", (preset_id,))
+    db.commit()
+    db.close()
+    return {"status": "ok"}

routers/profile.py

@@ -0,0 +1,36 @@
+"""JarvisChat routers - Profile."""
+from datetime import datetime, timezone
+from fastapi import APIRouter, HTTPException, Request
+from db import get_db
+from security import read_json_body, BODY_LIMIT_PROFILE_BYTES
+from config import MAX_PROFILE_CHARS, DEFAULT_PROFILE
+
+router = APIRouter()
+
+
+@router.get("/api/profile")
+async def get_profile():
+    db = get_db()
+    row = db.execute("SELECT content, updated_at FROM profile WHERE id = 1").fetchone()
+    db.close()
+    return ({"content": row["content"], "updated_at": row["updated_at"]} if row
+            else {"content": "", "updated_at": ""})
+
+
+@router.put("/api/profile")
+async def update_profile(request: Request):
+    body = await read_json_body(request, BODY_LIMIT_PROFILE_BYTES)
+    content = str(body.get("content", ""))
+    if len(content) > MAX_PROFILE_CHARS:
+        raise HTTPException(status_code=413, detail="Profile content is too long")
+    now = datetime.now(timezone.utc).isoformat()
+    db = get_db()
+    db.execute("UPDATE profile SET content = ?, updated_at = ? WHERE id = 1", (content, now))
+    db.commit()
+    db.close()
+    return {"status": "ok", "updated_at": now}
+
+
+@router.get("/api/profile/default")
+async def get_default_profile():
+    return {"content": DEFAULT_PROFILE}

routers/search_route.py

@@ -0,0 +1,108 @@
+"""JarvisChat routers - /api/search explicit search endpoint."""
+import json
+import logging
+import uuid
+from datetime import datetime, timezone
+
+import httpx
+from fastapi import APIRouter, HTTPException, Request
+from fastapi.responses import StreamingResponse
+
+from config import DEFAULT_MODEL, LLAMA_SERVER_BASE, MAX_SEARCH_QUERY_CHARS
+from db import get_db
+from search import query_searxng, format_search_results
+from routers.chat import parse_llama_stream_chunk
+from security import read_json_body, log_incident, BODY_LIMIT_CHAT_BYTES
+
+log = logging.getLogger("jarvischat")
+router = APIRouter()
+
+
+@router.post("/api/search")
+async def explicit_search(request: Request):
+    body = await read_json_body(request, BODY_LIMIT_CHAT_BYTES)
+    query = body.get("query", "").strip()
+    if len(query) > MAX_SEARCH_QUERY_CHARS:
+        raise HTTPException(status_code=413, detail="Search query is too long")
+    conv_id = body.get("conversation_id")
+    model = body.get("model", DEFAULT_MODEL)
+
+    if not query:
+        raise HTTPException(status_code=400, detail="Empty query")
+
+    db = get_db()
+    now = datetime.now(timezone.utc).isoformat()
+
+    if not conv_id:
+        conv_id = str(uuid.uuid4())
+        title = f"🔍 {query[:70]}..." if len(query) > 70 else f"🔍 {query}"
+        db.execute("INSERT INTO conversations (id, title, model, created_at, updated_at) VALUES (?, ?, ?, ?, ?)",
+                   (conv_id, title, model, now, now))
+    else:
+        db.execute("UPDATE conversations SET updated_at = ? WHERE id = ?", (now, conv_id))
+
+    db.execute("INSERT INTO messages (conversation_id, role, content, created_at) VALUES (?, ?, ?, ?)",
+               (conv_id, "user", f"🔍 {query}", now))
+    db.commit()
+    db.close()
+
+    async def stream_search():
+        yield f"data: {json.dumps({'conversation_id': conv_id, 'searching': True})}\n\n"
+
+        results = await query_searxng(query, max_results=5)
+
+        if not results:
+            error_msg = "No search results found."
+            yield f"data: {json.dumps({'token': error_msg, 'conversation_id': conv_id})}\n\n"
+            db2 = get_db()
+            db2.execute("INSERT INTO messages (conversation_id, role, content, created_at) VALUES (?, ?, ?, ?)",
+                        (conv_id, "assistant", error_msg, datetime.now(timezone.utc).isoformat()))
+            db2.commit()
+            db2.close()
+            yield f"data: {json.dumps({'done': True, 'conversation_id': conv_id})}\n\n"
+            return
+
+        yield f"data: {json.dumps({'search_results': len(results), 'conversation_id': conv_id})}\n\n"
+
+        search_context = format_search_results(results)
+        messages = [
+            {"role": "system", "content": f"You have access to current web data. Answer directly using ONLY the data below. Be concise. No apologies. No disclaimers.\n\n{search_context}"},
+            {"role": "user", "content": query},
+        ]
+
+        full_response = []
+        async with httpx.AsyncClient() as client:
+            try:
+                async with client.stream(
+                    "POST", f"{LLAMA_SERVER_BASE}/v1/chat/completions",
+                    json={"model": model, "messages": messages, "stream": True},
+                    timeout=httpx.Timeout(300.0, connect=10.0),
+                ) as resp:
+                    async for line in resp.aiter_lines():
+                        if line.strip():
+                            token, done, _ = parse_llama_stream_chunk(line)
+                            if token:
+                                full_response.append(token)
+                                yield f"data: {json.dumps({'token': token, 'conversation_id': conv_id})}\n\n"
+                            if done:
+                                break
+            except Exception as e:
+                incident_key = log_incident("search_summarization_stream",
+                                            message="Stream failure during explicit search summarization",
+                                            request=request, exc=e)
+                yield f"data: {json.dumps({'error': 'Search summarization could not complete right now.', 'error_key': incident_key})}\n\n"
+                return
+
+        summary = "".join(full_response)
+        saved_msg = f"{summary}\n\n---\n*🔍 Web search results*"
+
+        db2 = get_db()
+        db2.execute("INSERT INTO messages (conversation_id, role, content, created_at) VALUES (?, ?, ?, ?)",
+                    (conv_id, "assistant", saved_msg, datetime.now(timezone.utc).isoformat()))
+        db2.commit()
+        db2.close()
+
+        yield f"data: {json.dumps({'raw_results': results, 'conversation_id': conv_id})}\n\n"
+        yield f"data: {json.dumps({'done': True, 'conversation_id': conv_id, 'searched': True})}\n\n"
+
+    return StreamingResponse(stream_search(), media_type="text/event-stream")

routers/settings.py

@@ -0,0 +1,36 @@
+"""JarvisChat routers - Settings."""
+from fastapi import APIRouter, HTTPException, Request
+from db import get_db
+from security import read_json_body, BODY_LIMIT_DEFAULT_BYTES
+from config import MAX_SETTINGS_KEYS, MAX_SETTINGS_VALUE_CHARS, ALLOWED_SETTINGS_KEYS
+
+router = APIRouter()
+
+
+@router.get("/api/settings")
+async def get_settings():
+    db = get_db()
+    rows = db.execute("SELECT key, value FROM settings").fetchall()
+    db.close()
+    return {row["key"]: row["value"] for row in rows}
+
+
+@router.put("/api/settings")
+async def update_settings(request: Request):
+    body = await read_json_body(request, BODY_LIMIT_DEFAULT_BYTES)
+    if not isinstance(body, dict):
+        raise HTTPException(status_code=400, detail="Settings payload must be an object")
+    if len(body) > MAX_SETTINGS_KEYS:
+        raise HTTPException(status_code=413, detail="Too many settings in one request")
+    unknown_keys = sorted(key for key in body.keys() if str(key) not in ALLOWED_SETTINGS_KEYS)
+    if unknown_keys:
+        raise HTTPException(status_code=400, detail=f"Unknown setting key(s): {', '.join(unknown_keys)}")
+    db = get_db()
+    for key, value in body.items():
+        if len(str(key)) > 80 or len(str(value)) > MAX_SETTINGS_VALUE_CHARS:
+            db.close()
+            raise HTTPException(status_code=413, detail="Setting key/value too long")
+        db.execute("INSERT OR REPLACE INTO settings (key, value) VALUES (?, ?)", (key, str(value)))
+    db.commit()
+    db.close()
+    return {"status": "ok"}

routers/skills.py

@@ -0,0 +1,42 @@
+"""JarvisChat routers - Skills."""
+from fastapi import APIRouter, HTTPException, Request
+from db import get_db, get_setting, list_skills_with_state, set_skill_enabled
+from security import read_json_body, BODY_LIMIT_DEFAULT_BYTES
+from config import MAX_SKILL_KEY_CHARS, SKILLS_BY_KEY
+
+router = APIRouter()
+
+
+@router.get("/api/skills")
+async def list_skills():
+    db = get_db()
+    skills = list_skills_with_state(db)
+    db.close()
+    return {"skills": skills, "count": len(skills)}
+
+
+@router.get("/api/skills/active")
+async def list_active_skills():
+    db = get_db()
+    skills_enabled = get_setting(db, "skills_enabled", "true") == "true"
+    skills = list_skills_with_state(db)
+    db.close()
+    active = [s for s in skills if s["enabled"]] if skills_enabled else []
+    return {"skills": active, "count": len(active), "skills_enabled": skills_enabled}
+
+
+@router.put("/api/skills/{skill_key}")
+async def update_skill(skill_key: str, request: Request):
+    skill_key = skill_key.strip()
+    if len(skill_key) > MAX_SKILL_KEY_CHARS or skill_key not in SKILLS_BY_KEY:
+        raise HTTPException(status_code=404, detail="Unknown skill")
+    body = await read_json_body(request, BODY_LIMIT_DEFAULT_BYTES)
+    if "enabled" not in body or not isinstance(body.get("enabled"), bool):
+        raise HTTPException(status_code=400, detail="Field 'enabled' (boolean) is required")
+    db = get_db()
+    set_skill_enabled(db, skill_key, bool(body["enabled"]))
+    db.commit()
+    skills = list_skills_with_state(db)
+    db.close()
+    updated = next((s for s in skills if s["key"] == skill_key), None)
+    return {"status": "ok", "skill": updated}

search.py

@@ -0,0 +1,144 @@
+"""
+JarvisChat - SearXNG integration, perplexity scoring, refusal/hedge detection.
+"""
+import logging
+import math
+import re
+from urllib.parse import urlparse
+
+import httpx
+
+from config import SEARXNG_BASE, PERPLEXITY_THRESHOLD, REFUSAL_PATTERNS, HEDGE_PATTERNS
+
+log = logging.getLogger("jarvischat")
+
+
+def sanitize_outbound_url(url: str) -> str:
+    if not url:
+        return ""
+    candidate = url.strip()
+    parsed = urlparse(candidate)
+    if parsed.scheme.lower() in {"http", "https"} and parsed.netloc:
+        return candidate
+    return ""
+
+
+def calculate_perplexity(logprobs: list) -> float:
+    if not logprobs:
+        return 0.0
+    avg_logprob = sum(lp["logprob"] for lp in logprobs) / len(logprobs)
+    return math.exp(-avg_logprob)
+
+
+def is_uncertain(logprobs: list, threshold: float = PERPLEXITY_THRESHOLD) -> bool:
+    if not logprobs:
+        return False
+    perplexity = calculate_perplexity(logprobs)
+    log.info(f"Perplexity: {perplexity:.2f} (threshold: {threshold})")
+    return perplexity > threshold
+
+
+def is_refusal(text: str) -> bool:
+    match = REFUSAL_PATTERNS.search(text)
+    if match:
+        log.info(f"Refusal detected: '{match.group()}'")
+        return True
+    return False
+
+
+def clean_hedging(text: str) -> str:
+    cleaned = text
+    for pattern in HEDGE_PATTERNS:
+        cleaned = re.sub(pattern, "", cleaned, flags=re.IGNORECASE)
+    return cleaned.strip()
+
+
+def format_search_results(results: list) -> str:
+    if not results:
+        return ""
+    lines = ["[LIVE WEB DATA]\n"]
+    for i, r in enumerate(results, 1):
+        lines.append(f"{i}. {r['title']}")
+        if r["content"]:
+            lines.append(f"   {r['content']}")
+        lines.append("")
+    lines.append("\nAnswer directly using the data above. No apologies. No disclaimers. Just answer.")
+    return "\n".join(lines)
+
+
+def format_direct_answer(question: str, results: list) -> str:
+    if not results:
+        return "No search results found."
+    lines = ["Here's what I found:\n"]
+    for r in results[:3]:
+        lines.append(f"**{r['title']}**")
+        if r["content"]:
+            lines.append(f"{r['content']}")
+        lines.append("")
+    return "\n".join(lines).strip()
+
+
+def extract_search_query(user_message: str) -> str:
+    query = user_message.strip()
+    if re.search(r"temperature|weather", query, re.IGNORECASE):
+        query = re.sub(r"^what('?s| is) the ", "", query, flags=re.IGNORECASE) + " right now degrees"
+    if re.search(r"price|spot price", query, re.IGNORECASE):
+        query = re.sub(r"^(what('?s| is)|can you tell me) the ", "", query, flags=re.IGNORECASE) + " today USD"
+    query = re.sub(
+        r"^(what|who|where|when|why|how|is|are|can|could|would|should|do|does|did)\s+",
+        "", query, flags=re.IGNORECASE,
+    )
+    query = re.sub(r"[?!.]+$", "", query)
+    return query[:100].strip() or user_message[:100]
+
+
+async def query_searxng(query: str, max_results: int = 5) -> list:
+    log.info(f"Querying SearXNG: '{query}'")
+    async with httpx.AsyncClient() as client:
+        weather_match = re.search(
+            r"(?:weather|temperature|forecast)\s+(?:in\s+)?(.+?)(?:\s+right now|\s+today|\s+degrees)?$",
+            query, re.IGNORECASE,
+        )
+        if weather_match or "weather" in query.lower() or "temperature" in query.lower():
+            location = (
+                weather_match.group(1) if weather_match
+                else re.sub(r"(weather|temperature|forecast|right now|today|degrees)", "", query, flags=re.IGNORECASE).strip()
+            )
+            if location:
+                try:
+                    resp = await client.get(f"https://wttr.in/{location}?format=3", timeout=10.0,
+                                            headers={"User-Agent": "curl/7.68.0"})
+                    if resp.status_code == 200:
+                        return [{"title": "Current Weather",
+                                 "url": sanitize_outbound_url(f"https://wttr.in/{location}"),
+                                 "content": resp.text.strip()}]
+                except Exception as e:
+                    log.warning(f"wttr.in error: {e}")
+
+        try:
+            resp = await client.get(
+                f"{SEARXNG_BASE}/search",
+                params={"q": query, "format": "json", "categories": "general"},
+                timeout=10.0,
+            )
+            if resp.status_code == 200:
+                data = resp.json()
+                results = []
+                for answer in data.get("answers", []):
+                    results.append({"title": "Direct Answer", "url": "", "content": answer})
+                for box in data.get("infoboxes", []):
+                    content = box.get("content", "")
+                    if not content and box.get("attributes"):
+                        content = " | ".join([f"{a.get('label','')}: {a.get('value','')}" for a in box["attributes"]])
+                    results.append({
+                        "title": box.get("infobox", "Info"),
+                        "url": sanitize_outbound_url(box.get("urls", [{}])[0].get("url", "") if box.get("urls") else ""),
+                        "content": content,
+                    })
+                for r in data.get("results", [])[:max_results]:
+                    results.append({"title": r.get("title", ""), "url": sanitize_outbound_url(r.get("url", "")), "content": r.get("content", "")})
+                log.info(f"SearXNG returned {len(results)} results")
+                return results
+        except Exception as e:
+            log.error(f"SearXNG error: {e}")
+    return []

security.py

@@ -0,0 +1,175 @@
+"""
+JarvisChat - Security utilities.
+PIN hashing, audit logging, incident tracking, CSRF/origin checks,
+rate limiting, request helpers.
+"""
+import hashlib
+import hmac
+import json
+import logging
+import math
+import os
+import platform
+import time
+import uuid
+from collections import defaultdict, deque
+from datetime import datetime, timezone
+from threading import Lock
+from typing import Optional
+from urllib.parse import urlparse
+
+from fastapi import HTTPException, Request
+
+from config import (
+    ALLOWED_NETWORKS, TRUST_X_FORWARDED_FOR, TRUSTED_ORIGINS,
+    BODY_LIMIT_DEFAULT_BYTES, BODY_LIMIT_CHAT_BYTES, BODY_LIMIT_PROFILE_BYTES,
+    RATE_WINDOW_SECONDS, RL_LOGIN_PER_WINDOW, RL_CHAT_PER_WINDOW,
+    RL_SEARCH_PER_WINDOW, RL_STATS_PER_WINDOW, RL_WRITE_PER_WINDOW,
+    RL_DEFAULT_PER_WINDOW, VERSION,
+)
+
+import ipaddress
+
+log = logging.getLogger("jarvischat")
+
+SESSIONS: dict = {}
+PIN_ATTEMPTS: dict = {}
+RATE_EVENTS: dict = defaultdict(deque)
+SESSION_LOCK = Lock()
+RATE_LOCK = Lock()
+
+
+def hash_pin(pin: str, salt_hex: Optional[str] = None) -> tuple:
+    salt = bytes.fromhex(salt_hex) if salt_hex else os.urandom(16)
+    digest = hashlib.pbkdf2_hmac("sha256", pin.encode("utf-8"), salt, 200_000)
+    return salt.hex(), digest.hex()
+
+
+def audit_event(event: str, outcome: str, *, ip: str = "unknown", role: str = "none",
+                details: str = "", warning: bool = False) -> None:
+    payload = {"event": event, "outcome": outcome, "ip": ip, "role": role, "details": details[:300]}
+    msg = "AUDIT " + json.dumps(payload, separators=(",", ":"))
+    if warning:
+        log.warning(msg)
+    else:
+        log.info(msg)
+
+
+def create_incident_key() -> str:
+    ts = datetime.now(timezone.utc).strftime("%Y%m%d-%H%M%S")
+    return f"INC-{ts}-{uuid.uuid4().hex[:8].upper()}"
+
+
+def customer_error_envelope(message: str, incident_key: str) -> dict:
+    return {
+        "detail": message, "error_key": incident_key,
+        "error": {"message": message, "incident_key": incident_key,
+                  "support_hint": "Share this incident key for exact diagnostics."},
+    }
+
+
+def log_incident(event: str, *, message: str, request: Optional[Request] = None,
+                 exc: Optional[Exception] = None) -> str:
+    incident_key = create_incident_key()
+    payload = {
+        "event": event, "incident_key": incident_key, "message": message,
+        "app_version": VERSION, "pid": os.getpid(), "python": platform.python_version(),
+        "platform": platform.platform(),
+        "method": request.method if request else "",
+        "path": request.url.path if request else "",
+        "client_ip": get_client_ip(request) if request else "",
+    }
+    if exc:
+        log.exception("INCIDENT " + json.dumps(payload, separators=(",", ":")))
+    else:
+        log.error("INCIDENT " + json.dumps(payload, separators=(",", ":")))
+    return incident_key
+
+
+def get_client_ip(request: Request) -> str:
+    forwarded = request.headers.get("x-forwarded-for", "").strip()
+    if TRUST_X_FORWARDED_FOR and forwarded:
+        return forwarded.split(",")[0].strip()
+    if request.client and request.client.host:
+        return request.client.host
+    return "unknown"
+
+
+def is_ip_allowed(ip: str) -> bool:
+    normalized = ip.strip().lower()
+    if normalized in {"localhost", "testclient"}:
+        normalized = "127.0.0.1"
+    try:
+        ip_obj = ipaddress.ip_address(normalized)
+    except ValueError:
+        return False
+    for network in ALLOWED_NETWORKS:
+        if ip_obj in network:
+            return True
+    return False
+
+
+def request_body_limit(path: str) -> int:
+    if path in {"/api/chat", "/api/search"}:
+        return BODY_LIMIT_CHAT_BYTES
+    if path == "/api/profile":
+        return BODY_LIMIT_PROFILE_BYTES
+    return BODY_LIMIT_DEFAULT_BYTES
+
+
+def rate_policy(path: str, method: str, ip: str, sid: str) -> tuple:
+    identity = sid or ip
+    if path == "/api/auth/login":
+        return f"login:{ip}", RL_LOGIN_PER_WINDOW
+    if path == "/api/chat":
+        return f"chat:{identity}", RL_CHAT_PER_WINDOW
+    if path == "/api/search":
+        return f"search:{identity}", RL_SEARCH_PER_WINDOW
+    if path == "/api/stats":
+        return f"stats:{ip}", RL_STATS_PER_WINDOW
+    if method in {"POST", "PUT", "DELETE", "PATCH"}:
+        return f"write:{identity}", RL_WRITE_PER_WINDOW
+    return f"api:{identity}", RL_DEFAULT_PER_WINDOW
+
+
+def check_rate_limit(key: str, limit: int, window_seconds: int) -> tuple:
+    now_ts = time.time()
+    with RATE_LOCK:
+        bucket = RATE_EVENTS[key]
+        while bucket and bucket[0] <= (now_ts - window_seconds):
+            bucket.popleft()
+        if len(bucket) >= limit:
+            retry_after = max(1, int(math.ceil(window_seconds - (now_ts - bucket[0]))))
+            return False, retry_after
+        bucket.append(now_ts)
+    return True, 0
+
+
+def origin_allowed(request: Request) -> bool:
+    host = request.headers.get("host", "").strip()
+    expected_origin = f"{request.url.scheme}://{host}".rstrip("/") if host else ""
+    origin = request.headers.get("origin", "").strip().rstrip("/")
+    referer = request.headers.get("referer", "").strip()
+    if origin:
+        return origin == expected_origin or origin in TRUSTED_ORIGINS
+    if referer:
+        parsed = urlparse(referer)
+        ref_origin = f"{parsed.scheme}://{parsed.netloc}".rstrip("/")
+        return ref_origin == expected_origin or ref_origin in TRUSTED_ORIGINS
+    return True
+
+
+def is_state_changing(method: str) -> bool:
+    return method in {"POST", "PUT", "DELETE", "PATCH"}
+
+
+async def read_json_body(request: Request, max_bytes: int) -> dict:
+    raw = await request.body()
+    if len(raw) > max_bytes:
+        raise HTTPException(status_code=413, detail="Request payload too large")
+    if not raw:
+        return {}
+    try:
+        return json.loads(raw.decode("utf-8"))
+    except Exception:
+        raise HTTPException(status_code=400, detail="Invalid JSON payload")

templates/index.html

@@ -227,7 +227,7 @@ body { font-family: var(--font-body); background: var(--bg-primary); color: var(
 
 <aside class="sidebar" id="sidebar">
     <div class="sidebar-header">
-        <img class="logo" src="/static/logo.jpg" alt="JarvisChat Logo" onerror="this.style.display='none'">
+        <img class="logo" src="/static/logo.png" alt="JarvisChat Logo" onerror="this.style.display='none'">
         <h1>⚡ JarvisChat {{ version }}</h1>
         <div class="subtitle">🦙 local coding companion</div>
         <div class="btn-row">