afce0bd3e5441750494b061a6fb68fa0f9dbe8e0
Switch from external Gemini API (3072 dims, $0.15/1M tokens) to local Ollama mxbai-embed-large (1024 dims, free) for cost savings and HNSW index support. Changes: - Updated embeddings.ts: model 'mxbai-embed-large', API URL fixed - Updated migration 015: vector(1024) with HNSW index - Regenerated 268 tool_docs embeddings with new model Benefits: - Free embeddings (no API costs) - HNSW index enabled (1024 < 2000 dim limit) - Fast similarity search (O(log n) vs O(n)) - No external API dependency Trade-offs: - 5% quality loss (MTEB 64.68 vs ~70 Gemini) - Uses local compute (1.2GB RAM, <1s per embedding) Task: CF-251 Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Task MCP Server
Exposes task management tools via Model Context Protocol. Uses PostgreSQL with pgvector for semantic search.
Requirements
- SSH tunnel to services:
ssh -L 5432:localhost:5432 services -N & - PostgreSQL with pgvector on services (CI stack
postgrescontainer) - CI postgres must be running:
ssh services "cd /opt/docker/ci && docker compose up -d postgres"
Configuration
Add to ~/.claude/settings.json under mcpServers:
{
"task-mcp": {
"command": "node",
"args": ["/path/to/task-mcp/dist/index.js"],
"env": {
"DB_HOST": "localhost",
"DB_PORT": "5432",
"DB_NAME": "agiliton",
"DB_USER": "agiliton",
"DB_PASSWORD": "<from /opt/docker/ci/.env>",
"LLM_API_URL": "https://llm.agiliton.cloud",
"LLM_API_KEY": "sk-master-..."
}
}
}
Tools
| Tool | Description |
|---|---|
task_add |
Create task with auto-generated ID and embedding |
task_list |
List tasks with filters (project, status, type, priority) |
task_show |
Show task details including checklist and dependencies |
task_close |
Mark task as completed |
task_update |
Update task fields |
task_similar |
Find semantically similar tasks using pgvector |
task_context |
Get related tasks for current work context |
task_link |
Create dependency between tasks |
task_checklist_add |
Add checklist item to task |
task_checklist_toggle |
Toggle checklist item |
Build
npm install
npm run build
Development
npm run dev # Run with tsx (no build)
Description
Languages
TypeScript
73.8%
JavaScript
19.5%
PLpgSQL
6.3%
Shell
0.3%