qdrant-fabric
ry-ops/qdrant-fabric
Model Context Protocol server for Qdrant - Part of Infrastructure as a Fabric ecosystem. Provides 30+ tools for vector database operations, semantic search, and AI memory management.
Instalação
npx convez-ai ry-ops/qdrant-fabric1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.
Métricas
README
A tradução para o português deste README ainda está na fila.
Qdrant Fabric
A comprehensive Model Context Protocol (MCP) server providing access to both Qdrant Cloud Management and Database APIs as part of the Infrastructure as a Fabric ecosystem.
Features
- 30 Database Tools: Complete Phase 1 implementation with all core database operations
- Full MCP Integration: Works seamlessly with Claude Desktop and other MCP clients
- Type-Safe: Full Pydantic validation and type hints
- Async-First: Built on modern async Python patterns
- Fabric Integration: Seamlessly integrates with AIANA and n8n-fabric
- Comprehensive: Collections, Points, Search, Payload, Health, Vectors, and Index operations
Infrastructure Fabric
The qdrant-fabric MCP server sits at the heart of the infrastructure, connecting:
- Local Qdrant (localhost:6333) - Your development instance with collections
- Qdrant Cloud - Managed clusters, backups, and production infrastructure
- AIANA - AI Assistant with semantic memory (
aiana_memoriescollection) - n8n-fabric - Workflow automation with vector indexing (
n8n_workflowscollection)
This hybrid architecture enables seamless local development while providing full cloud management capabilities.
Installation
pip install qdrant-fabric
Quick Start
Configure Claude Desktop
Add to your Claude Desktop config (macOS: ~/Library/Application Support/Claude/claude_desktop_config.json):
{
"mcpServers": {
"qdrant-fabric": {
"command": "uv",
"args": [
"run",
"--directory",
"/path/to/qdrant-fabric",
"python",
"-m",
"qdrant_mcp"
],
"env": {
"QDRANT_API_KEY": "your-database-api-key",
"QDRANT_URL": "http://localhost:6333"
}
}
}
}
Or use Python directly:
{
"mcpServers": {
"qdrant-fabric": {
"command": "python",
"args": ["-m", "qdrant_mcp"],
"env": {
"QDRANT_API_KEY": "your-database-api-key",
"QDRANT_URL": "http://localhost:6333"
}
}
}
}
Available Tools (v0.0.4)
All 30 Phase 1 database tools are now available:
Collections Management (6 tools)
qdrant_db_collections_list- List all collectionsqdrant_db_collections_get- Get collection detailsqdrant_db_collections_create- Create new collectionqdrant_db_collections_delete- Delete collectionqdrant_db_collections_update- Update collection configurationqdrant_db_collections_exists- Check if collection exists
Points Operations (7 tools)
qdrant_db_points_upsert- Insert or update pointsqdrant_db_points_get- Retrieve multiple points by IDqdrant_db_points_get_single- Get single point by IDqdrant_db_points_delete- Delete pointsqdrant_db_points_count- Count points with optional filterqdrant_db_points_scroll- Scroll through pointsqdrant_db_points_batch- Batch point operations
Vector Search (4 tools)
qdrant_db_points_search- Vector similarity searchqdrant_db_points_search_batch- Batch search queriesqdrant_db_points_recommend- Recommendation based on examplesqdrant_db_points_recommend_batch- Batch recommendations
Payload Management (4 tools)
qdrant_db_payload_set- Set payload (merge with existing)qdrant_db_payload_overwrite- Overwrite payload (replace)qdrant_db_payload_delete- Delete specific payload fieldsqdrant_db_payload_clear- Clear all payload
Health Checks (5 tools)
qdrant_db_health_root- Version and build informationqdrant_db_health_check- Health checkqdrant_db_health_liveness- Liveness probeqdrant_db_health_readiness- Readiness probeqdrant_db_health_metrics- Prometheus metrics
Vector Operations (2 tools)
qdrant_db_vectors_update- Update vectors for existing pointsqdrant_db_vectors_delete- Delete named vectors
Index Management (2 tools)
qdrant_db_index_create- Create field index for faster filteringqdrant_db_index_delete- Delete field index
API Coverage
Phase 1: Core Database Operations ✅ Complete (v0.0.4)
- ✅ Collections Management (6 tools)
- ✅ Points Operations (7 tools)
- ✅ Vector Search (4 tools)
- ✅ Payload Management (4 tools)
- ✅ Health Checks (5 tools)
- ✅ Vector Operations (2 tools)
- ✅ Index Management (2 tools)
Total: 30 tools
Coming Soon
- Phase 2: Cloud Management Essentials
- Phase 3: Advanced Search & Discovery
- Phase 4: Backup & Recovery
- See API_COVERAGE_PLAN.md for full roadmap
Configuration
Environment Variables
Cloud Management API:
QDRANT_CLOUD_API_KEY- Bearer token for Qdrant Cloud APIQDRANT_CLOUD_URL- Cloud API base URL (default:https://cloud.qdrant.io)
Database API:
QDRANT_API_KEY- API key for database access (optional for local instances)QDRANT_URL- Database URL (e.g.,http://localhost:6333orhttps://xyz.qdrant.io)
Note: Cloud Management API tools are coming in Phase 2. Currently, only Database API tools are available.
Development
Setup
git clone https://github.com/yourusername/qdrant-fabric.git
cd qdrant-fabric
pip install -e ".[dev]"
Run Tests
pytest
Code Quality
black src/ tests/
ruff check src/ tests/
mypy src/
Architecture
qdrant-fabric/
├── src/qdrant_mcp/
│ ├── server.py # MCP server entrypoint
│ ├── config.py # Configuration management
│ ├── cloud/ # Cloud Management API tools
│ └── database/ # Database API tools
└── tests/ # Test suite
Documentation
- API Coverage Plan - Complete roadmap
- Cloud API Surface - All gRPC methods
- Database API Operations - All REST endpoints
Contributing
Contributions welcome! Please read our contributing guidelines and submit PRs.
License
MIT License - see LICENSE file for details