F
licenseC
qualityD
maintenanceProvides RAG tools with local vector retrieval and web fallback using Firecrawl, enabling document ingestion and querying through MCP stdio transport.
3
-
kb-agent-platform
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., "@kb-agent-platformWhat are the key points from the Q3 roadmap?"
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.
KB-Agent-Platform — Enterprise Knowledge Base Q&A Platform (RAG + Multi-Agent + MCP)
In one sentence: an enterprise knowledge base Agent based on RAG and LangGraph Supervisor, with SSE streaming Q&A with citation tracing, dual MCP entry points, and one-click deployment via docker compose.
Architecture
Frontend Vue3 ──SSE──> FastAPI ──> Supervisor (LLM routing) ──┬─> KB Agent ──> Qdrant (4096-dim)
└─> Chat Agent
MCP Host (Cursor/Claude) ──stdio──> MCP Server ──> same Qdrant
Related MCP server: tero-mcp-lite
Quick Start
https://github.com/embracingTheFuture/kb-agent-platform.git && cd kb-agent-platform
cp .env.example .env # fill in your key
docker compose up --build -d
docker compose run --rm api python task2_pipeline.py # ingest data
Screenshots
Quality: Evaluation Set Regression
6 test cases covering routing correctness + citation integrity, reproducible inside the container.
ID | First 15 characters | Expected | Actual | Passed |
ID: 1 | First 15 characters: Phase 2 schedule and main project positioning. | Expected: KB | Actual: KB | PASS |
ID: 2 | First 15 characters: What Day16 in the roadmap is for | Expected: KB | Actual: KB | PASS |
ID: 3 | First 15 characters: Introduce yourself in one sentence. | Expected: CHAT | Actual: CHAT | PASS |
ID: 4 | First 15 characters: Help me write a poem about autumn. | Expected: CHAT | Actual: CHAT | PASS |
ID: 5 | First 15 characters: How many medals did Team China win at the 2024 Olympics | Expected: CHAT | Actual: CHAT | PASS |
ID: 6 | First 15 characters: | Expected: CHAT | Actual: CHAT | PASS |
6 test cases, 6 passed, 0 failed |
Technical Highlights
Handwritten ReAct core (tool registry + while loop, no framework)
Self-developed chunking-vectorization-retrieval pipeline, Qdrant cosine similarity
Citation tracing: numbered context + [i] annotations + citations structure
Supervisor multi-agent routing + conversation memory compression (rolling summary)
MCP Server dual entry points / SSE streaming / evaluation set regression
This server cannot be deployed
Maintenance
ActivityMaintained
ResponsivenessNo issues
Related MCP Connectors
Query any docs site via MCP. Submit a URL, ask questions, get cited answers.
- docs2mcpOAuthcom.docs2mcp
Query your own PDFs and documents from any MCP client. Every answer cites the page it came from.
Query SEC EDGAR filings, XBRL financials, and company data through MCP. STDIO & Streamable HTTP.
Serve a folder of Markdown notes as an MCP server: hybrid search, reading, and sourced answers.
Related MCP Servers
- AlicenseBqualityAmaintenanceA lightweight MCP server that enables querying a project's corpus (docs, decisions, issues, skills) with cited answers and typed refusals via stdio JSON-RPC 2.0.9MIT
- AlicenseNot gradedqualityBmaintenanceProvides a stdio MCP bridge for coding agents to query and record engineering knowledge locally, preserving debugging history, failed attempts, and verified solutions.2 npmMIT
- AlicenseNot gradedqualityCmaintenanceEnables retrieval-augmented generation over a local markdown corpus, allowing grounded, cited answers via an MCP tool or CLI.6 npmMIT