Qdrant Vector DB MCP

Run vector operations against Qdrant through the MAD MCP broker with configurable ingest workflows for documents and textbooks.

Authentication

• Header: Authorization

• Format: Bearer mad_live_***

• Configure Qdrant URL + API key in Integrations before using broker tools.

• Use qdrant-ingest-textbook for large textbook workflows (async job pipeline).

• Use qdrant-ingest-document for non-textbook document/content ingestion.

Catalog Coverage

• Total tools: 57
• GET tools: 3
• POST tools: 54
• Snapshot version: snapshot-20260330T033151Z
• Snapshot captured at: 2026-03-30T03:31:51.722Z
• Docs generated at: 2026-03-30T04:36:10.795Z

GET Tools

• Get Ingest Status — Get status, stage, and diagnostics for an async textbook ingest job.
• List Collections — Return collections currently available for the configured cluster.
• Semantic Search — Run similarity search with optional metadata filters.

POST Tools

• Audit Memories — Audit memory payloads for missing fields and duplicates.
• Backfill Memory Contract — Backfill missing memory contract fields for existing points.
• Bulk Patch — Apply metadata/payload patches to points by id or filter.
• Cache Memory — Store short-term memory with a TTL in a cache collection.
• Cancel Ingest Job — Request cancellation for a running textbook ingest job.
• Cancel Job — Cancel a running job.
• Collection Aliases — List aliases for a specific collection.
• Collection Cluster Info — Get cluster info for a collection.
• Collection Exists — Check if a collection exists.
• Collection Info — Get collection details including vectors and payload schema.
• Collection Payload Schema — Get payload schema for a collection.
• Collection Stats — Get basic collection statistics (points, segments, status).
• Collection Vectors — List vector names and sizes for a collection.
• Count Points — Count points matching an optional filter.
• Create Snapshot — Create a collection snapshot (admin-only, confirm required).
• Dedupe Memories — Find and optionally delete duplicate memories.
• Delete By Filter — Delete points by filter (confirm required).
• Delete Document — Delete all chunks for a document by doc_id (confirm required).
• Delete Points — Delete points by id (confirm required).
• Ensure Payload Indexes — Ensure expected payload indexes exist for a collection.
• Expire Memories — Expire memories by expires_at_ts (optional archive).
• Expire Short Term — Expire short-term memories by expires_at_ts.
• Find Near Duplicates — Find near-duplicate points using vector similarity.
• Find Short Term — Search the short-term memory cache collection.
• Get Points — Retrieve points by id.
• Get Vector Name — Resolve the vector name used by this MCP server.
• Health Check — Run health checks against Qdrant and embedding clients.
• Ingest Document — Ingest general documents/content (document workflow) from source URL or inline text.
• Ingest Textbook (Async) — Submit a source_url textbook ingestion job and return immediately with job_id.
• Ingest With Validation — Store memory with contract validation and quarantine support.
• Job Logs — Fetch recent logs for a submitted job.
• Job Progress — Get progress for a submitted job.
• Job Result — Fetch the result for a completed job.
• Job Status — Check status for a submitted job.
• Link Memories — Link memories via related_ids and optional associations.
• List Aliases — List all collection aliases.
• List Full Snapshots — List full cluster snapshots.
• List Points — List points with pagination (scroll).
• List Shard Snapshots — List snapshots for a specific shard.
• List Snapshots — List snapshots for a collection.
• Merge Duplicates — Merge duplicate points into a canonical point.
• Metrics Snapshot — Snapshot collection stats and index coverage metrics.
• Optimizer Status — Get optimizer config and index coverage for a collection.
• Patch Payload — Patch payload metadata for a point.
• Promote Short Term — Promote short-term memories into the long-term collection.
• Recommend Memories — Recommend memories using positive/negative example ids.
• Reembed Points — Re-embed points when embedding version changes.
• Restore Snapshot — Restore a collection snapshot (admin-only, confirm required).
• Store Vectorized Text — Chunk and store text with metadata in a target collection.
• Submit Job — Submit a long-running housekeeping job.
• Tag Memories — Append or replace labels for a set of points.
• Update Optimizer Config — Update optimizer config (admin; confirm + dry_run=false required).
• Update Point — Update an existing point (re-embeds content).
• Validate Memory — Validate memory contract fields before ingest.