A
licenseNot graded
qualityB
maintenanceEnables hybrid document search (BM25 and dense) over a configurable corpus via MCP tools, returning passages and sources for AI agents to cite in answers.
MIT
atlas-mcp-doc-search
by bragabruno
Provides BM25 keyword search over the doc_chunks index as part of hybrid document retrieval, combining sparse signals with dense vector search.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@atlas-mcp-doc-searchsearch for articles on Kubernetes security best practices"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
atlas-mcp-doc-search
MCP server that exposes hybrid document search over the Atlas ingested corpus. Built with the official Python mcp SDK (FastMCP server interface, Streamable HTTP transport), Python 3.12 asyncio, and deployed as a standalone K8s service on AKS.
Tool Contract
doc_search
doc_search(query: str, k: int = 8) -> {chunks: [{id, text, source_id, score}]}Parameter | Type | Default | Description |
|
| required | Natural-language search query |
|
|
| Number of top-ranked chunks to return |
Returns a list of up to k chunks, each with:
Field | Type | Description |
|
| Unique chunk identifier |
|
| Raw chunk text |
|
| Identifier of the source document |
|
| Reciprocal Rank Fusion (RRF) fused score |
Related MCP server: RAG Knowledge Base MCP Server
Hybrid Retrieval Approach
The server implements HYBRID retrieval — combining sparse (BM25) and dense (vector) signals and fusing them with Reciprocal Rank Fusion (RRF):
Query embedding — the raw query string is sent to the Atlas gateway's
/v1/embeddingsendpoint to produce a dense vector.Parallel retrieval
Elasticsearch BM25 keyword search over the
doc_chunksindex.Qdrant vector similarity search over collection
doc_chunks(payload fields:source_id,doc_id,chunk_idx,text).
Fusion — both ranked lists are merged with RRF to produce a single ranked list.
Return — top
kresults are returned with their fused scores.
Dependencies
Dependency | Role |
| MCP server framework |
Elasticsearch | BM25 keyword retrieval over |
Qdrant | Dense vector retrieval over collection |
Atlas gateway | Query embedding generation |
Diagrams
Related
atlas-docs — document ingestion pipeline that populates the corpus
atlas-mcp-citations — citation verification MCP server
This server cannot be deployed
Maintenance
ActivityInactive
ResponsivenessNo issues
Related MCP Connectors
Search your knowledge bases from any AI assistant using hybrid RAG.
Semantic search across 50,000+ food recipes with hybrid retrieval and reranking.
Cloud or self-hosted knowledge for AI agents: hybrid search, reranking, GraphRAG, scoped MCP tools.
Search source collections your team vetted. Every passage cites its URL and fetch date.
Related MCP Servers
- FlicenseNot gradedqualityCmaintenanceEnables searching a knowledge base and asking grounded questions with hybrid retrieval, reranking, and cited answers.-
- AlicenseNot gradedqualityBmaintenanceProvides hybrid retrieval (dense + BM25 + RRF) with collection-based isolation and document ingestion for private knowledge access via MCP.MIT
- FlicenseNot gradedqualityCmaintenanceEnables querying a hybrid-retrieval RAG pipeline (dense + BM25) over ingested PDF documents, returning answers generated by Gemini.-