qdrant-mcp
qdrant-mcp
Servidor MCP para la ingesta de documentos y búsqueda semántica en Qdrant.
Descripción general
qdrant-mcp proporciona herramientas para:
ingerir documentos locales en una colección de Qdrant
generar embeddings con OpenAI
ejecutar búsquedas vectoriales con filtros de metadatos opcionales
Related MCP server: mcp-server-qdrant
Características
ingest_documentsconvierte archivos como
docx,pptxypdfa Markdown mediante MarkItDowndivide el contenido en fragmentos usando
chunk_sizeyoverlap_ratioincrusta fragmentos con OpenAI Embeddings (
text-embedding-3-smallpor defecto)realiza upsert del texto del fragmento y los metadatos en Qdrant
search_documentsincrusta el texto de consulta con la misma API de embeddings
recupera las
kmejores coincidencias de Qdrantadmite filtrado por
categoryypath
Requisitos
Python 3.11+
uvQdrant (por ejemplo,
http://localhost:6333)OPENAI_API_KEY
Configuración
uv syncEjecutar dentro de la CLI de Codex
[mcp_servers.qdrant-mcp]
command = "uv"
args = ["run", "qdrant-mcp"]
cwd = "/sandbox/qdrant-mcp"
env = {
OPENAI_API_KEY = "sk-...",
QDRANT_URL = "http://127.0.0.1:6333",
QDRANT_API_KEY = "QDRANT_API_KEY",
QDRANT_COLLECTION = "codex_collection",
CHUNK_HEADER_MODEL = "gpt-5.4-mini"
}Pruebas
Establezca OPENAI_API_KEY, QDRANT_URL y QDRANT_API_KEY en .env, luego ejecute:
uv run python -m unittest tests/integration/test_qdrant_integration.pyHerramientas MCP
ingest_documents
Parámetros:
paths: list[str]category: strchunk_size: int = 1200overlap_ratio: float = 0.15embedding_model: str = "text-embedding-3-small"chunk_header_mode: Literal["enabled", "disabled"] = "enabled"
Retorna:
collectionembedding_modelingested_filesingested_pointsfailed_files
search_documents
Parámetros:
query: strtop_k: int = 5category: str | None = Nonepath: str | None = Noneembedding_model: str = "text-embedding-3-small"
Retorna:
collectionembedding_modelquerycountresults(score,path,category,chunk_index,text)
delete_documents_by_path
Parámetros:
path: strcategory: str | None = None
Retorna:
collectionpathcategorystatusoperation_id
list_category
Parámetros:
limit: int = 100
Retorna:
collectioncountcategories
list_path
Parámetros:
category: strlimit: int = 1000
Retorna:
collectioncategorycountpaths
Notas
Si la colección de destino no existe, se crea automáticamente en la primera ingesta.
Si los índices de carga útil para
categoryypathno existen, se crean durante la ingesta.Por defecto, la ingesta antepone un
Chunk-Headergenerado (máximo 64 caracteres), derivado de los primeros 4096 bytes, a cada fragmento.El modelo de Chunk-Header se lee de
CHUNK_HEADER_MODELcuandochunk_header_modeestá enenabled(por defecto:gpt-5.4-mini).El nombre de la colección se configura solo a través de
QDRANT_COLLECTION(no mediante parámetros de la herramienta MCP).Con
text-embedding-3-small, el tamaño del vector es1536.
Licencia
Ver LICENSE.
This server cannot be installed
Maintenance
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Related MCP Servers
- AlicenseAqualityDmaintenanceLocal-first RAG indexing and semantic search MCP server. Enables document retrieval and context-aware queries using local embedding models.39MIT
- Alicense-qualityCmaintenanceMCP server for Qdrant vector database with local BERT embeddings. Enables semantic search and vector storage operations through natural language.MIT
- Alicense-qualityDmaintenanceMCP server that searches documents in Qdrant using embeddings from LMStudio. Takes a text query, converts it to a vector via LMStudio's OpenAI-compatible API, and performs semantic search in Qdrant.13ISC
- Flicense-qualityCmaintenanceMCP server for retrieving answers from local PDFs using RAG with FAISS and OpenAI.
Related MCP Connectors
Remote ChromaDB vector database MCP server with streamable HTTP transport
Agent-native MCP server over the public saagarpatel.dev corpus. Read-only, stateless.
MCP server for AgentDocs (agentdocs.eu): read, search, write, comment on & share Markdown docs.
Latest Blog Posts
- Who's Calling? MCP Hosts Are an Identity Blind Spot (And the Spec Knows It)By Om-Shree-0709 on .mcpAgent IdentityOAuth 2.1
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/msstnk/qdrant-mcp'
If you have feedback or need assistance with the MCP directory API, please join our Discord server