Skip to content
Deutsch

MCP server in namespace io.github

io.github.sfrangulov/minirag-mcp · Status: active

The source provides no name of its own for this entry. The identifier therefore appears on this page as a coordinate, not as the heading.

If you find your own account name here: write informally to German.saas@web.de, stating the identifier concerned. No reason has to be given. How to object

Registry name
io.github.sfrangulov/minirag-mcp
Package coordinate
pypi:minirag-mcp
Version
0.6.0
Listed in the registry
2026-08-09
Changed in the registry
2026-08-09
First seen by tracevero
2026-08-10
Vendor's website
https://github.com/sfrangulov/minirag-mcp#readme

Local-first RAG MCP server: hybrid search over a folder of your own documents

Vendor's own description, untranslated and unverified

Evidence for this page: The values on this page come from several collections. Each row therefore states its own origin. What the confidence levels mean

Measured values

Measured values
Credentials requiredno
Raw declaration – Required secret variables (raw): confirmed absent
Raw declaration – Required secret headers (raw): confirmed absent
Source: MCP-Register · collected on 2026-08-10 · derived
Execution locationlocal
Raw declaration – Transports (raw): stdio
Source: MCP-Register · collected on 2026-08-10 · derived
Path argument presentyes
Raw declaration – Path arguments (raw): confirmed absent
Raw declaration – Path environment variables (raw): BASE_DIR, CACHE_DIR, DB_PATH
Source: MCP-Register · collected on 2026-08-16 · derived
Repository URL listedyes
Raw declaration – Repository (raw): https://github.com/sfrangulov/minirag-mcp
Source: MCP-Register · collected on 2026-08-10 · derived
Version0.6.0
Source: MCP-Register · collected on 2026-08-10 · self-declared

Position within the holdings

How many of the 22,700 published entries carry the same measured value. A reference figure, not an assessment.

All 4 values together are carried by 171 of 22,700 entries.

This entry appears in these selections

The source's raw values23

Evidence for this page: The values on this page come from several collections. Each row therefore states its own origin.

Raw values
Environment variables (raw)BASE_DIR, BASE_DIRS, DB_PATH, CACHE_DIR, MODEL_NAME, MAX_FILE_SIZE, CHUNK_TOKEN_BUDGET, RAG_HYBRID_WEIGHT, RAG_GROUPING, RAG_MAX_DISTANCE, RAG_MAX_FILES, RAG_INSTRUCTIONS_APPEND, ALLOW_PRIVATE_URLS
Source: MCP-Register · collected on 2026-08-10 · self-declared
Required secret variables (raw)confirmed absent
Source: MCP-Register · collected on 2026-08-10 · self-declared
Transports (raw)stdio
Source: MCP-Register · collected on 2026-08-10 · self-declared
Path arguments (raw)confirmed absent
Source: MCP-Register · collected on 2026-08-10 · self-declared
Repository (raw)https://github.com/sfrangulov/minirag-mcp
Source: MCP-Register · collected on 2026-08-10 · self-declared
Package registries (raw)pypi
Source: MCP-Register · collected on 2026-08-10 · self-declared
Required secret headers (raw)confirmed absent
Source: MCP-Register · collected on 2026-08-16 · self-declared
Path environment variables (raw)BASE_DIR, CACHE_DIR, DB_PATH
Source: MCP-Register · collected on 2026-08-16 · self-declared
Remote URLs (raw)confirmed absent
Source: MCP-Register · collected on 2026-08-16 · self-declared
Remote hosts (raw)confirmed absent
Source: MCP-Register · collected on 2026-08-16 · self-declared
Registry status message (raw)confirmed absent
Source: MCP-Register · collected on 2026-08-16 · self-declared
First listed in the registry (raw)2026-08-09
Source: MCP-Register · collected on 2026-08-16 · self-declared
Last changed in the registry (raw)2026-08-09
Source: MCP-Register · collected on 2026-08-16 · self-declared
Schema version of the raw record (raw)https://static.modelcontextprotocol.io/schemas/2025-12-11/server.schema.json
Source: MCP-Register · collected on 2026-08-16 · self-declared
Delivery form (raw)paket
Source: MCP-Register · collected on 2026-08-16 · self-declared
Repository platform (raw)github
Source: MCP-Register · collected on 2026-08-16 · self-declared
Repository subfolder (raw)confirmed absent
Source: MCP-Register · collected on 2026-08-16 · self-declared
Package identifiers (raw)minirag-mcp
Source: MCP-Register · collected on 2026-08-16 · self-declared
Package versions (raw)0.6.0
Source: MCP-Register · collected on 2026-08-16 · self-declared
Runtime hints (raw)uvx
Source: MCP-Register · collected on 2026-08-16 · self-declared
Environment variable formats (raw)boolean, filepath, number, string
Source: MCP-Register · collected on 2026-08-16 · self-declared
Environment variable descriptions (raw)ALLOW_PRIVATE_URLS=Let ingest_url fetch hosts resolving to loopback, link-local, private, reserved or unspecified addresses. Off by default. · BASE_DIR=One document root; also the security boundary for file access. Defaults to the process working directory. Ignored when BASE_DIRS is set. · BASE_DIRS=JSON array of document roots, e.g. ["/docs/a", "/docs/b"]. Takes precedence over BASE_DIR. An invalid value is a hard configuration error. · CACHE_DIR=Embedding model cache. Defaults to the platform user cache dir, so the ~220 MB model is downloaded once and shared. · CHUNK_TOKEN_BUDGET=Retrieval-unit size in the embedding model's own tokens. Range 16-128; the ceiling is the model's trained sequence length. · DB_PATH=LanceDB index directory. Defaults to <first root>/.minirag/lancedb, so each corpus gets its own index. · MAX_FILE_SIZE=Per-file size limit in bytes, enforced before parsing. · MODEL_NAME=fastembed model id. Changing it makes existing vectors incompatible with new queries; pair with a new DB_PATH or a full re-ingest. · RAG_GROUPING=Result grouping filter. 'similar' keeps only the closest group; 'related' also keeps the next one. Unset means no grouping filter. · RAG_HYBRID_WEIGHT=Keyword weight in the weighted RRF fusion, range 0.0-1.0. 0 is vector-only; higher values raise the BM25 contribution. · RAG_INSTRUCTIONS_APPEND=Extra paragraph appended to the instructions the server hands the client at connect time, for corpus-specific guidance. · RAG_MAX_DISTANCE=Drop results whose vector distance exceeds this value. Lower is stricter. Unset means no distance filter. · RAG_MAX_FILES=Keep chunks from at most this many best-scoring files. Unset means no per-file filter.
Source: MCP-Register · collected on 2026-08-16 · self-declared
Icon formats (raw)confirmed absent
Source: MCP-Register · collected on 2026-08-16 · self-declared

Changes

Changes
2026-08-16Path argument present: false → true
2026-08-16Last changed in the registry (raw): – → 2026-08-09
2026-08-16First listed in the registry (raw): – → 2026-08-09
2026-08-16Registry status message (raw): – → –
2026-08-16Schema version of the raw record (raw): – → https://static.modelcontextprotocol.io/schemas/2025-12-11/server.schema.json
2026-08-16Icon formats (raw): – → –
2026-08-16Environment variable descriptions (raw): – → ALLOW_PRIVATE_URLS=Let ingest_url fetch hosts resolving to loopback, link-local, private, reserved or unspecified addresses. Off by default. · BASE_DIR=One document root; also the security boundary for file access. Defaults to the process working directory. Ignored when BASE_DIRS is set. · BASE_DIRS=JSON array of document roots, e.g. ["/docs/a", "/docs/b"]. Takes precedence over BASE_DIR. An invalid value is a hard configuration error. · CACHE_DIR=Embedding model cache. Defaults to the platform user cache dir, so the ~220 MB model is downloaded once and shared. · CHUNK_TOKEN_BUDGET=Retrieval-unit size in the embedding model's own tokens. Range 16-128; the ceiling is the model's trained sequence length. · DB_PATH=LanceDB index directory. Defaults to <first root>/.minirag/lancedb, so each corpus gets its own index. · MAX_FILE_SIZE=Per-file size limit in bytes, enforced before parsing. · MODEL_NAME=fastembed model id. Changing it makes existing vectors incompatible with new queries; pair with a new DB_PATH or a full re-ingest. · RAG_GROUPING=Result grouping filter. 'similar' keeps only the closest group; 'related' also keeps the next one. Unset means no grouping filter. · RAG_HYBRID_WEIGHT=Keyword weight in the weighted RRF fusion, range 0.0-1.0. 0 is vector-only; higher values raise the BM25 contribution. · RAG_INSTRUCTIONS_APPEND=Extra paragraph appended to the instructions the server hands the client at connect time, for corpus-specific guidance. · RAG_MAX_DISTANCE=Drop results whose vector distance exceeds this value. Lower is stricter. Unset means no distance filter. · RAG_MAX_FILES=Keep chunks from at most this many best-scoring files. Unset means no per-file filter.
2026-08-16Environment variable formats (raw): – → boolean, filepath, number, string
2026-08-16Delivery form (raw): – → paket
2026-08-16Remote hosts (raw): – → –

These are the 10 most recent changes to this entry. Full history

tracevero · https://tracevero.com/en/mcp/io-github-sfrangulov-minirag-mcp