Knowledge Management · Developer Tools
shervinemp/CodebaseMCP
A RAG system based on Python code analysis that parses code structure via AST and stores it in a Weaviate vector database, providing code queries, natural language Q&A and visualization, with support for multi-codebase management and dependency analysis.
LLM_PROVIDERLLM_API_KEYGENERATE_LLM_DESCRIPTIONSLLM_CONCURRENCYGENERATION_MODEL_NAMEEMBEDDING_MODEL_NAMEWEAVIATE_HOSTWEAVIATE_PORTWEAVIATE_GRPC_PORTSEMANTIC_SEARCH_LIMITSEMANTIC_SEARCH_DISTANCEWATCHER_POLLING_INTERVALMCP server that scans Python codebases with AST, stores them in Weaviate, and answers questions via pluggable LLM providers (Gemini, OpenAI).
AST scanning — functions, classes, imports, calls, variables, decorators, signatures, docstrings, cross-references.
Hybrid search — Weaviate vector + BM25 keyword search. Unenriched elements are still findable by name/code content.
DAG-ordered enrichment — callees enriched before callers, so refinement sees real callee descriptions.
RAG Q&A — ask_question retrieves relevant code via hybrid search and synthesises answers with an LLM.
File watching — rescans and re-enriches on file changes, DAG-ordered.
Worker pool — enrichment uses LLM_CONCURRENCY workers with a queue instead of N asyncio tasks.
Codebase dependencies — relationships between codebases; queries span dependencies in parallel.
Call graph visualization — MermaidJS from stored cross-references.
scan_codebase
Scan, upload, DAG-enrich, summarise, start watcher
list_codebases
Registered codebases with status
select_codebase
Set active context (stops prior watcher)
delete_codebase
Remove codebase + tenant + registry entry
find_element
Search by name, file, type across active + dependencies
get_details
Full properties for a UUID
analyze_snippet
Find elements related to a code snippet
ask_question
RAG (with optional include_dependencies)
trigger_llm_processing
Queue enrichment (DAG-ordered, worker pool)
regenerate_summary
Re-run summary without rescanning
start_watcher
Manual watcher control
add_codebase_dependency
Dependency graph management
Details on this page are taken from the project's README. Open README
1Panel-dev/MaxKB
MaxKB = Max Knowledge Brain, it is an open-source platform for building enterprise-grade agents. MaxKB integrates Retrieval-Augmented Generation (RAG) pipelines, supports robust workflows, and provides advanced MCP tool-use capabilities.
bytedance/UI-TARS-desktop
TARS is ByteDance's multimodal AI agent stack, shipping two projects: Agent TARS (a CLI and Web UI agent built on MCP) and UI-TARS-desktop (a desktop GUI agent).
supermemoryai/supermemory
State-of-the-art memory and context engine for AI.
topoteretes/cognee
Cognee - The Free Open-Source AI Memory Platform for Agents.
makenotion/notion-mcp-server
For the best experience, use Remote Notion MCP, our official hosted MCP server. This repository is a separate, self-hosted MCP server implementation that is no longer actively maintained or supported. Please use Remote Notion MCP instead.
cyberagiinc/devdocs
[!WARNING] DevDocs Status: Not publicly maintained. Enhanced internal version at CyberAGI — public release coming soon. If you have any questions please reach out to info@cyberagi.ai.