Replace main with rebuilt AI-Todo

.env.example

@@ -1,133 +1,16 @@
-# =============================================================================
-# AI-Todo configuration
-# =============================================================================
-#
-# Copy this file to `~/.agentmemory/.env` (run `node dist/cli.mjs init`) and
-# uncomment the lines you want. Every line is OFF by default — AI-Todo runs
-# out of the box: it extracts To-Dos from your agent sessions with a
-# deterministic rules extractor, no LLM key, no API auth.
-#
-# Run `node dist/cli.mjs doctor` to verify the daemon reads the env you expect.
-# Env var NAMES (AGENTMEMORY_* / LANGEXTRACT_* / III_*) are the wire contract
-# and are kept stable — only this file's framing/order is AI-Todo-oriented.
-
-# -----------------------------------------------------------------------------
-# 1. To-Do extraction model  ← AI-Todo's headline config
-# -----------------------------------------------------------------------------
-#
-# To use an LLM extractor instead of the default rules extractor, install the
-# Python deps then set an OpenAI-compatible model below:
-#   python3 -m pip install -r requirements-langextract.txt
-#   python3 -m pip install socksio  # only when your network uses a SOCKS proxy
-#
-# First-run setup (`node dist/cli.mjs --reset` from source, or
-# `agentmemory-lab --reset` when installed) seeds these for the model you pick.
-# Only the openai-compatible branch is wired (Novita default / OpenAI / OpenRouter).
-# AGENTMEMORY_TODO_EXTRACTOR=langextract
-# LANGEXTRACT_PROVIDER=openai
-# LANGEXTRACT_MODEL=deepseek/deepseek-v4-flash   # Novita; OpenAI: gpt-4o-mini; OpenRouter: deepseek/deepseek-chat
-# LANGEXTRACT_BASE_URL=https://api.novita.ai/openai/v1  # OpenAI: https://api.openai.com/v1 · OpenRouter: https://openrouter.ai/api/v1
-# LANGEXTRACT_API_KEY=<runtime secret>
-# LANGEXTRACT_THINKING_DEPTH=medium              # reasoning_effort; leave empty for non-reasoning models (e.g. gpt-4o-mini)
-# LANGEXTRACT_PYTHON=/path/to/python
-# AGENTMEMORY_TODO_EXTRACT_TIMEOUT_MS=120000
-# AGENTMEMORY_TODO_EXTRACT_SINCE_DAYS=7                 # only extract sessions from the last N days (primary scope control)
-# AGENTMEMORY_TODO_EXTRACT_MAX_INTERACTIONS_PER_SESSION=10  # per session, keep at most M most-recent interaction records (one user request → agent reply)
-
-# -----------------------------------------------------------------------------
-# 2. Auth & security (web UI / REST)
-# -----------------------------------------------------------------------------
-#
-# Bearer-token auth for the REST API + viewer. Without a secret, REST endpoints
-# are open on loopback. Set this when you expose the daemon beyond loopback.
-# AGENTMEMORY_SECRET=your-secret-here
-
-# -----------------------------------------------------------------------------
-# 3. Runtime / ports / engine pin
-# -----------------------------------------------------------------------------
-
-# AGENTMEMORY_URL=http://localhost:3111          # REST base URL — honored by status, doctor, MCP shim
-# AGENTMEMORY_VIEWER_URL=http://localhost:3114   # Override the viewer URL printed by `node dist/cli.mjs status` / `agentmemory-lab status`
-# AGENTMEMORY_TOOLS=all                          # core (default) | all — surface exposed to MCP clients
-# AGENTMEMORY_SLOTS=memory                       # Comma-separated plugin slot names the CLI should claim
-# AGENTMEMORY_EXPORT_ROOT=~/agentmemory-backup   # Default destination used by export functions
-# AGENTMEMORY_DEBUG=1                             # Trace MCP shim probe + standalone fallback decisions to stderr
-# AGENTMEMORY_FORCE_PROXY=1                       # Skip the MCP shim livez probe and trust AGENTMEMORY_URL
-# AGENTMEMORY_PROBE_TIMEOUT_MS=2000              # MCP shim livez probe timeout
-# STANDALONE_MCP=1                               # MCP shim only — bypass the worker, run @agentmemory/mcp in-process
-# STANDALONE_PERSIST_PATH=~/.agentmemory/local.db  # Path used by the standalone MCP shim's local fallback store
-# III_REST_PORT=3111                             # REST API port (also affects viewer at +2)
-# III_STREAMS_PORT=3112                          # Streams API port
-# III_ENGINE_URL=ws://localhost:49134            # iii-engine WebSocket URL (used by the worker)
-# AGENTMEMORY_III_VERSION=0.11.2                 # iii-engine pin. v0.11.6 changes the sandbox model; only override after migrating manually.
-
-# =============================================================================
-# Advanced: legacy memory engine (inherited from agentmemory; OFF by default, #138)
-# =============================================================================
-#
-# These power the upstream persistent-memory layer (compression, consolidation,
-# embeddings, knowledge graph, snapshots, team sharing, bridges). AI-Todo does
-# NOT need them for To-Do extraction; they remain for users who want the memory
-# layer. Names unchanged.
-
-# --- LLM provider (compress / summarise / consolidate) — pick ONE -------------
-# Detection order: OPENAI_API_KEY → MINIMAX_API_KEY → ANTHROPIC_API_KEY →
-# GEMINI_API_KEY → OPENROUTER_API_KEY → noop (zero-LLM synthetic compression).
-# OPENAI_API_KEY=sk-...                          # Also used for OpenAI-compatible embeddings
-# OPENAI_BASE_URL=https://api.openai.com         # Override for OpenAI-compatible providers
-# ANTHROPIC_API_KEY=sk-ant-...
-# ANTHROPIC_MODEL=claude-sonnet-4-20250514       # Default Anthropic model
-# ANTHROPIC_BASE_URL=https://api.anthropic.com   # Override for Anthropic-compatible proxies / Azure AI Foundry
-# GEMINI_API_KEY=...                             # GEMINI_API_KEY takes precedence over GOOGLE_API_KEY
-# GOOGLE_API_KEY=...                             # Alias for GEMINI_API_KEY when set alone
-# GEMINI_MODEL=gemini-2.5-flash                  # Default Gemini model
-# OPENROUTER_API_KEY=sk-or-...
-# OPENROUTER_MODEL=anthropic/claude-sonnet-4-20250514
-# MINIMAX_API_KEY=...
-# MINIMAX_MODEL=MiniMax-M2.7
-# MAX_TOKENS=4096                                # Cap LLM completion tokens for compression / summarise calls
-# AGENTMEMORY_LLM_TIMEOUT_MS=60000               # Outbound LLM / embedding timeout (default 60 000 ms)
-# AGENTMEMORY_ALLOW_AGENT_SDK=true               # Opt-in Claude-subscription fallback (off by default)
-# FALLBACK_PROVIDERS=anthropic,gemini            # Chain tried after the primary provider errors
-
-# --- Embeddings (override via EMBEDDING_PROVIDER) -----------------------------
-# Detection order: EMBEDDING_PROVIDER → GEMINI → OPENAI → VOYAGE → COHERE →
-# OPENROUTER → local (Xenova/all-MiniLM-L6-v2, 384-dim).
-# EMBEDDING_PROVIDER=local                       # local | openai | voyage | cohere | gemini | openrouter
-# VOYAGE_API_KEY=pa-...                          # Optimised for code embeddings
-# COHERE_API_KEY=...                             # General-purpose embeddings
-# OPENAI_EMBEDDING_MODEL=text-embedding-3-small  # When EMBEDDING_PROVIDER=openai
-# OPENAI_EMBEDDING_DIMENSIONS=1536               # Required when the model is not in the known-models table
-# OPENROUTER_EMBEDDING_MODEL=openai/text-embedding-3-small  # When EMBEDDING_PROVIDER=openrouter
-
-# --- Search tuning -----------------------------------------------------------
-# BM25_WEIGHT=0.4                                # Hybrid search weight for BM25 leg
-# VECTOR_WEIGHT=0.6                              # Hybrid search weight for vector leg
-# AGENTMEMORY_GRAPH_WEIGHT=0.2                   # Graph traversal bonus on smart-search ranking
-# TOKEN_BUDGET=2000                              # Max tokens injected via mem::context per session
-# MAX_OBS_PER_SESSION=500                        # Per-session observation cap before consolidation kicks in
-# SUMMARIZE_CHUNK_SIZE=400                       # Chunk threshold for map-reduce summarize of large sessions
-# SUMMARIZE_CHUNK_CONCURRENCY=6                  # Parallel chunk LLM calls during chunked summarize
-
-# --- Behaviour flags ---------------------------------------------------------
-# AGENTMEMORY_AUTO_COMPRESS=true                 # LLM compression on every observation batch (needs a provider key). Default off.
-# AGENTMEMORY_INJECT_CONTEXT=true                # Inject recalled memories back into agent prompts (#143). Default off.
-# CONSOLIDATION_ENABLED=true                     # Run the 4-tier consolidation pipeline. Default off.
-# CONSOLIDATION_DECAY_DAYS=30                    # Age (days) after which non-reinforced memories decay
-# GRAPH_EXTRACTION_ENABLED=true                  # Extract concept-graph edges on remember
-# GRAPH_EXTRACTION_BATCH_SIZE=8                  # Memories per graph-extraction batch
-# AGENTMEMORY_REFLECT=true                       # Periodically auto-synthesize lessons from memories
-# AGENTMEMORY_DROP_STALE_INDEX=true              # Drop on-disk BM25 / vector index on startup if dim guard fires (#248)
-# AGENTMEMORY_IMAGE_EMBEDDINGS=true              # Enable image embeddings when an image provider is present (experimental)
-
-# --- Snapshots / team / bridges ----------------------------------------------
-# SNAPSHOT_ENABLED=true
-# SNAPSHOT_DIR=~/.agentmemory/snapshots
-# SNAPSHOT_INTERVAL=3600                         # Seconds between snapshots
-# TEAM_MODE=shared                               # Scope memories to (TEAM_ID, USER_ID) tuples
-# TEAM_ID=acme
-# USER_ID=rohit
-# CLAUDE_MEMORY_BRIDGE=true                      # Mirror compressed memories into Claude Code's CLAUDE.md
-# CLAUDE_PROJECT_PATH=/path/to/your/project      # Required when CLAUDE_MEMORY_BRIDGE=true
-# CLAUDE_MEMORY_LINE_BUDGET=200                  # Lines of memory CLAUDE.md should hold
-# OBSIDIAN_AUTO_EXPORT=true                      # Auto-export memories to an Obsidian vault on every consolidation
+# AI-Todo local configuration example
+AI_TODO_CODEX_HOME=~/.codex
+AI_TODO_CLAUDE_HOME=~/.claude/projects
+AI_TODO_LLM_ENABLED=true
+AI_TODO_LLM_PROVIDER=openai
+AI_TODO_LLM_MODEL=deepseek/deepseek-v4-flash
+AI_TODO_LLM_ENDPOINT=https://api.novita.ai/openai/v1
+AI_TODO_LLM_THINKING_DEPTH=medium
+AI_TODO_LLM_TIMEOUT_MS=120000
+AI_TODO_ORGANIZE_SINCE_DAYS=7
+AI_TODO_ORGANIZE_MAX_INTERACTIONS_PER_SESSION=10
+AI_TODO_ORGANIZE_MAX_SESSIONS=16
+AI_TODO_ORGANIZE_MAX_OBSERVATIONS_PER_SESSION=40
+
+# Do not commit real keys. Run `ai-todo init --api-key ...` or save in Settings.
+# AI_TODO_LLM_API_KEY=