aichallenge-rag
Integrates with OpenAI-compatible embedding APIs to generate vector embeddings for document indexing and search queries.
This server does not currently have a deployable release on Glama. Follow the repository's setup instructions to install and configure it for your MCP client, including any required credentials and local dependencies.
aichallenge-rag
Standalone RAG index / search service for AIChallenge and any client that can call HTTP or Streamable HTTP MCP.
Vector backend: Qdrant — embedded on disk by default (RAG_DATA_DIR/qdrant), or a remote server via QDRANT_URL.
Index a folder of docs (or upload files), retrieve chunks with cosine similarity + metadata filters, optionally filter + heuristic-rerank, then feed the context into your LLM. This process never calls a chat model for answers — only embeddings (API, local, or fake).
You run | AIChallenge / client sees |
Qdrant index + search on your machine (or stand) | HTTP |
Embedded store under | Bearer token only |
Origin in the monorepo: apps/rag. This repo is the public, self-contained copy.
Features
Qdrant vector store — COSINE dense vectors, payload filters (
scope,owner_id, …)Corpus index — walk
RAG_CORPUS_DIR(md/txt/pdf with text layer, common source suffixes)Two chunk strategies —
fixedwindow vsstructural(headings / files)User uploads —
POST /v1/documents/uploadwith optionalowner_idscopeRetrieval modes —
raw·filtered(min score) ·full(rewrite + filter + heuristic rerank)/v1/ask— returns a readycontext/prompt_suffixblock (no LLM inside)MCP tools —
rag_stats,rag_search,rag_indexon/mcpEmbeddings — OpenAI-compatible API, optional
sentence-transformers, or deterministicfake
Related MCP server: RAG MCP Server
5-minute local run
git clone https://github.com/ArtemKyslicyn/aichallenge-rag.git
cd aichallenge-rag
cp .env.example .env
# optional: set RAG_SHARED_TOKEN to a long random string
uv sync
uv run python -m aichallenge_ragHealth: http://127.0.0.1:18766/health
OpenAPI: http://127.0.0.1:18766/docs
Stats (
backend=qdrant):GET /v1/stats
With the bundled corpus/sample.md and EMBEDDING_PROVIDER=fake:
curl -sS -X POST http://127.0.0.1:18766/v1/search \
-H 'Content-Type: application/json' \
-d '{"query":"chunk strategies","mode":"full"}'Tests:
uv run pytestQdrant modes
Mode | Config | When |
Embedded (default) |
| Local demos, Compose stand, CI |
Server |
| Shared / larger corpora |
# optional remote
docker run -d --name qdrant -p 6333:6333 qdrant/qdrant
export QDRANT_URL=http://127.0.0.1:6333Collection name: QDRANT_COLLECTION (default aichallenge_rag).
Connect to AIChallenge (Guest MCP)
Run the service on loopback (
RAG_HOST=127.0.0.1).Set the same long token in
.envasRAG_SHARED_TOKEN.Tunnel HTTPS to the process, e.g.
cloudflared tunnel --url http://127.0.0.1:18766.On the site: Profile → Подключения (Guest MCP)
URL:
https://<tunnel-host>/mcpToken: same as
RAG_SHARED_TOKEN
Full walkthrough: docs/connect-aichallenge.md.
HTTP API (summary)
Method | Path | Purpose |
|
| Liveness (no auth) |
|
| Chunk / vector / Qdrant backend |
|
| Rebuild corpus ( |
|
| Index if empty / rebuild if inconsistent |
|
| Add plain-text document (JSON) |
|
| Upload file (multipart) |
|
| List docs ( |
|
| Retrieve hits + scores |
|
| Search + |
|
| Flip embed provider / chunk strategy |
Details and examples: docs/api.md.
When RAG_SHARED_TOKEN is set, send Authorization: Bearer <token> on all routes except /health, /, /docs, /openapi.json.
Docker
docker compose up --buildBinds loopback only 127.0.0.1:18766. Never publish on public :443 / :8443 without your own edge plan.
Optional local embeddings: uv sync --extra local, then PATCH /v1/settings {"local_embeddings": true}.
Docs
Doc | Topic |
Qdrant pipeline, modes | |
HTTP + MCP contracts | |
Tunnel + Guest MCP | |
Tokens, bind, risks |
Env (names only)
See .env.example. Important names:
RAG_SHARED_TOKEN— Bearer for HTTP + MCP (empty = open local/dev)RAG_CORPUS_DIR/RAG_DATA_DIR— corpus + Qdrant path / metaQDRANT_URL/QDRANT_COLLECTION— remote vs embeddedEMBEDDING_PROVIDER—api|local|fakeLLM_API_KEY/ROUTERAI_KEY— for API embeddingsRAG_MODE/RAG_MIN_SCORE/RAG_TOP_K_*— retrieval defaults
Requirements
Python 3.12+
uv recommended
qdrant-client(pulled byuv sync)Optional: embedding API key, or
uv sync --extra localOptional: cloudflared / ngrok for Guest MCP
Security
Whoever has URL + Bearer can reindex, search, and upload into your index. Keep bind on loopback; only expose via a tunnel you control. See docs/security.md.
License
MIT — see LICENSE.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
This server cannot be deployed
Maintenance
Related MCP Connectors
Cloud or self-hosted knowledge for AI agents: hybrid search, reranking, GraphRAG, scoped MCP tools.
Serve a folder of Markdown notes as an MCP server: hybrid search, reading, and sourced answers.
Ingest, manage, and retrieve documents for RAG-powered AI applications
Agent-native MCP server over the public saagarpatel.dev corpus. Read-only, stateless.
Related MCP Servers
- AlicenseAqualityDmaintenanceLocal-first RAG indexing and semantic search MCP server. Enables document retrieval and context-aware queries using local embedding models.38 npmMIT
- FlicenseNot gradedqualityDmaintenanceIndexes PDF documents into Qdrant and exposes semantic search as MCP tools, enabling RAG-based interactions with your documents.-
- AlicenseNot gradedqualityBmaintenanceLocal RAG over a directory, served as an MCP tool plus CLI, with incremental indexing, zero-config defaults, and offline CPU-only operation.MIT
- AlicenseNot gradedqualityCmaintenanceSelf-hosted RAG MCP server that enables semantic and hybrid search over document corpora using local FAISS embeddings, with tools for indexing, retrieving chunks or full documents, uploading files, and managing multiple corpora via MCP.MIT