Codebase Memory MCP
High-performance code intelligence MCP server for AI coding agents. Indexes a codebase into a queryable knowledge graph in milliseconds, with 14 MCP tools spanning structural search, call-chain tracing, impact analysis, dead-code detection, and Cypher queries. Single static C binary, 66 languages via tree-sitter, zero runtime dependencies.
“The most active server in this batch and one of the highest-momentum projects in the entire directory. 100 commits in the last 30 days, 31 releases shipped in 9 weeks since first commit, and a peer-style arXiv preprint backing the architecture. Indexes the Linux kernel (28M LOC, 75K files) in 3 minutes, answers structural queries in under 1ms, and ships as a single static C binary with zero dependencies. SLSA Level 3 attestation bundles ship with every release. Designed for AI coding agents that need persistent code intelligence across sessions: tree-sitter AST parsing across 66 languages plus LSP-style hybrid type resolution for Go, C, and C++. The arXiv preprint (2603.27277) reports the trade-off: 83% answer quality versus 92% for a file-exploration agent, achieved at 10x fewer tokens across 31 real-world repositories. The pitch is the cost-quality curve, slightly lower raw quality at dramatically lower token cost, which matters when agents work in large repositories session after session.”
INSTALL THIS SERVER
{
"mcpServers": {
"codebase-memory": {
"command": "codebase-memory-mcp"
}
}
}
{
"mcpServers": {
"codebase-memory": {
"command": "codebase-memory-mcp"
}
}
}
{
"mcpServers": {
"codebase-memory": {
"command": "codebase-memory-mcp"
}
}
}
{
"mcpServers": {
"codebase-memory": {
"command": "codebase-memory-mcp"
}
}
}
{
"mcpServers": {
"codebase-memory": {
"command": "codebase-memory-mcp"
}
}
}
9 SELECTED TOOLS
OUR ASSESSMENT
High-performance code intelligence MCP server. MIT-licensed, distributed as a signed static C binary with SLSA Level 3 attestation bundles per release. 1,929 GitHub stars, 216 forks, 31 releases shipped, 100 commits in the last 30 days. Auto-detects 11 coding agents on install. Single static binary with zero runtime dependencies. arXiv preprint (2603.27277) publishes the architecture and benchmarks: 83% answer quality vs 92% for file-exploration baseline, at 10x fewer tokens across 31 repositories.
- 100 commits in the last 30 days with 31 releases shipped; the most actively developed server in this batch.
- Single static binary with zero runtime dependencies (no Docker, no Node, no Python).
- 66 languages via vendored tree-sitter grammars compiled into the binary.
- LSP-style hybrid type resolution for Go, C, and C++.
- Auto-detects 11 coding agents (Claude Code, Codex CLI, Gemini CLI, Zed, OpenCode, Antigravity, Aider, KiloCode, VS Code, OpenClaw, Kiro) on install.
- SLSA Level 3 attestation bundles per release; binaries are signed and checksummed.
- Optional 3D graph UI for human exploration of the knowledge graph.
- Infrastructure-as-code indexing: Dockerfiles, Kubernetes manifests, Kustomize overlays as graph nodes with cross-references.
- Local-only processing; codebase data stays on the operator machine.
- C codebase requires C build toolchain for source-level audits; most operators consume the binary.
- 9-week project age means edge-case behaviour is still maturing despite the high commit count.
- arXiv preprint is preprint status; the published benchmarks reflect the maintainer evaluation harness.
All processing happens locally; codebase data stays on the host machine. Release binaries are signed and checksummed; SLSA Level 3 attestation bundles ship alongside each binary asset. The install command writes to agent configuration files (Claude Code, VS Code, etc.) and pre-tool hooks; review the install script before running on locked-down hosts. OpenSSF Scorecard score: 5.8/10. Source is fully open at github.com/DeusData/codebase-memory-mcp for audit.
AI coding agents working in large repositories where token cost matters, teams operating polyglot codebases (66 supported languages), and engineers wanting structural impact analysis (caller graphs, HTTP route dependencies, blast-radius queries).
TECHNICAL DETAILS
RECORDED ADOPTION METRICS
Saved directory snapshots. Record last checked MAY 3, 2026. The review date above does not refresh these figures.
// Recorded note2,021 stars on the DeusData/codebase-memory-mcp repo. 76 commits on main in the last 30 days. Audit (2026-05-03) refreshed star count from 1,929 to 2,021.
// Recorded noteFirst-ranked in ai-ml category; near top-5 globally by recent activity and editorial weight.
SOURCES & RECORDED CHECKS
These are the directory listings and check details held in this record. Client compatibility, directory listings and installation testing are separate signals. Recorded client names describe the scope of a check; they do not establish a security assessment or a recurring test schedule.
The same server, 5 different lenses. We reconcile these signals into our editorial score, which is why our number sometimes diverges from a directory-aggregate star count.
| Source | Their rating | Their star count | Their downloads | Last synced |
|---|---|---|---|---|
| AutomationSwitch This page | 4.6editorial | 2,021 | — | MAY 3, 2026 |
| PulseMCP | — unrated | unavailable | unavailable | APR 29, 2026 |
| MCP.so | — unrated | unavailable | unavailable | APR 29, 2026 |
| Glama | — unrated | unavailable | unavailable | APR 29, 2026 |
| Smithery | — unrated | unavailable | unavailable | APR 29, 2026 |
| Official MCP Registry | — unrated | unavailable | unavailable | APR 29, 2026 |
// Counts are directory-reported; we don't adjust them. Discrepancies usually come from different snapshot times or star-caching.
OTHER AI / ML MCP SERVERS
Cognee MCP
Knowledge graph plus vector memory engine for AI agents, exposed as an MCP server with V2 session-aware memory tools (remember, recall, forget, improve) and classic V1 ingestion pipelines (cognify, codify). Three transports: stdio, SSE, Streamable HTTP. 16,965 GitHub stars, Apache-2.0.
HeyGen Hyperframes
Write HTML, render video. Built for agents. HeyGen's framework that turns HTML templates into video output via MCP. 16,400 GitHub stars and 100 commits on main in the last 30 days.
Arize Phoenix MCP
LLM observability platform exposing prompts, projects, traces, spans, sessions, datasets, and experiments through MCP. Published to npm as @arizeai/phoenix-mcp, current 4.0.8 (2026-04-29). 9,496 stars on parent monorepo, Elastic License 2.0.
Qdrant MCP Server
Official Qdrant vector database MCP server. Acts as a semantic memory layer on top of Qdrant: store information with metadata, retrieve via similarity search. Two tools, very small surface area, exceptionally maintained by the Qdrant team. Configurable embedding provider (fastembed default), supports remote and local Qdrant clusters.
Mastra
TypeScript framework for building AI-powered applications and agents, with built-in MCP server support. From the team behind Gatsby. 23,680 GitHub stars and 100 commits on main in the last 30 days.
Amazon Bedrock AgentCore MCP
Official AWS Labs MCP server for Amazon Bedrock AgentCore: agent runtime, memory, gateway, identity, and observability. Tools fetch curated AgentCore documentation and surface deployment guides for runtime, memory, and gateway resources. Apache-2.0 within awslabs/mcp monorepo (8,924 parent stars).
DISCUSS YOUR
MCP REQUIREMENTS.
Evaluating a server, scoping an internal deployment, or working out whether MCP is the right fit at all. Start the conversation and we will point you at the right piece of the ecosystem.