Qdrant Fabric

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_memories collection)
• n8n-fabric - Workflow automation with vector indexing (n8n_workflows collection)

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 collections
• qdrant_db_collections_get - Get collection details
• qdrant_db_collections_create - Create new collection
• qdrant_db_collections_delete - Delete collection
• qdrant_db_collections_update - Update collection configuration
• qdrant_db_collections_exists - Check if collection exists

Points Operations (7 tools)

• qdrant_db_points_upsert - Insert or update points
• qdrant_db_points_get - Retrieve multiple points by ID
• qdrant_db_points_get_single - Get single point by ID
• qdrant_db_points_delete - Delete points
• qdrant_db_points_count - Count points with optional filter
• qdrant_db_points_scroll - Scroll through points
• qdrant_db_points_batch - Batch point operations

Vector Search (4 tools)

• qdrant_db_points_search - Vector similarity search
• qdrant_db_points_search_batch - Batch search queries
• qdrant_db_points_recommend - Recommendation based on examples
• qdrant_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 fields
• qdrant_db_payload_clear - Clear all payload

Health Checks (5 tools)

• qdrant_db_health_root - Version and build information
• qdrant_db_health_check - Health check
• qdrant_db_health_liveness - Liveness probe
• qdrant_db_health_readiness - Readiness probe
• qdrant_db_health_metrics - Prometheus metrics

Vector Operations (2 tools)

• qdrant_db_vectors_update - Update vectors for existing points
• qdrant_db_vectors_delete - Delete named vectors

Index Management (2 tools)

• qdrant_db_index_create - Create field index for faster filtering
• qdrant_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 [APICOVERAGEPLAN.md](docs/APICOVERAGEPLAN.md) for full roadmap

Configuration

Environment Variables

Cloud Management API:

• QDRANT_CLOUD_API_KEY - Bearer token for Qdrant Cloud API
• QDRANT_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:6333 or https://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](docs/APICOVERAGEPLAN.md) - Complete roadmap
• [Cloud API Surface](docs/cloudapisurface.txt) - All gRPC methods
• [Database API Operations](docs/databaseapioperations.txt) - All REST endpoints

Contributing

Contributions welcome! Please read our contributing guidelines and submit PRs.

License

MIT License - see LICENSE file for details

Links

• Qdrant Documentation
• Qdrant Cloud Console
• Model Context Protocol

Source & license

This open-source MCP server is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: ry-ops
• Source: ry-ops/qdrant-fabric
• License: MIT

Install and usage instructions live in the source repository linked above.