codanna
Codanna is a local-first code intelligence server written in Rust that provides semantic search, call-graph analysis, and relationship tracking for codebases. It integrates with AI agents via the MCP protocol to help coding assistants understand code structure without relying on external APIs.
Key facts
Objective fields from the source. Values we can't verify are shown as “Unknown” rather than guessed.
| Field | Value |
|---|---|
| Repository | bartolli/codanna |
| Owner | bartolli |
| Primary language | Rust |
| License | Apache-2.0 — OSI-approved |
| Stars | 706 |
| Forks | 63 |
| Open issues | 10 |
| Latest release | v0.9.23 (2026-07-04) |
| Last updated | 2026-07-04 |
| Source | https://github.com/bartolli/codanna |
What codanna is
A Rust-based MCP server that indexes code symbols, parses multiple languages (14+ supported), performs semantic search via embeddings, and exposes call graphs and dependency analysis through stdio/HTTP transports. Achieves sub-10ms lookups and parses 75,000+ symbols per second.
Get the codanna source
Clone the repository and explore it locally.
git clone https://github.com/bartolli/codanna.gitcd codanna# follow the project's README for install & configurationNeed it deployed, integrated, or customized instead? DEV.co ships production installs.
Best use cases
Implementation considerations
- Requires initial indexing step (codanna init, codanna index) before use; index size and refresh strategy must be planned based on codebase churn.
- Embedding model (~150MB) downloads on first execution; ensure network access and storage available during initialization.
- MCP integration setup varies by client (Claude, Cursor, Windsurf); follow integration guides for each platform to avoid configuration errors.
- Language detection and symbol parsing depend on file extensions and AST analysis; test on representative code samples to confirm accuracy.
- Search quality depends on semantic embedding quality; custom profiles may be needed for domain-specific code patterns or terminology.
When to avoid it — and what to weigh
- Windows Production Deployments — Windows support is marked as experimental. If you rely on Windows servers or CI/CD, test thoroughly or consider alternatives.
- Minimal Infrastructure or Embedded Scenarios — Requires ~150MB for the embedding model on first run. Not suitable for resource-constrained environments or embedded systems.
- Need for Vendor Support or SLAs — This is a community open-source project. No commercial support, SLA, or liability guarantees. Enterprise projects requiring support agreements should evaluate commercial alternatives.
- Polyglot Codebases with Exotic Languages — 14+ languages are supported, but esoteric or proprietary languages may not be indexed. Verify your primary languages are on the supported list.
License & commercial use
Apache License 2.0 (SPDX: Apache-2.0). Permissive OSI-approved license. Allows commercial use, modification, and distribution with attribution and notice file retention required. See LICENSE and NOTICE files in repo.
Apache 2.0 permits commercial use, closed-source modification, and distribution. Attribution and NOTICE file inclusion are mandatory. No patent indemnification or explicit warranty. For commercial deployments, review NOTICE file requirements and consult legal if integrating with proprietary systems. No commercial support is evident from the repo.
DEV.co evaluation signals
Editorial assessment — not user reviews. Directional, with an explicit confidence level.
| Signal | Assessment |
|---|---|
| Maintenance | Active |
| Documentation | Strong |
| License clarity | Clear |
| Deployment complexity | Low |
| DEV.co fit | Good |
| Assessment confidence | High |
Local-first architecture reduces attack surface vs. cloud services. Runs as local process; standard OS process isolation applies. No assertion made about input validation, dependency vulnerability scanning, or cryptographic hardening. Embedding model and SSL/TLS usage (libssl-dev dependency) noted but not detailed. For sensitive code indexing, review data residency and access controls on the local server.
Alternatives to consider
Tree-sitter + LSP (e.g., rust-analyzer, pylsp)
Standard language server protocol offers symbol search and diagnostics but typically slower semantic understanding and less natural-language query capability. Better for IDE integration; less suitable for AI agent training loops.
Anthropic's Codebase Context (Claude API native)
Cloud-hosted, does not require local indexing, but introduces latency, API costs, and data residency concerns. Tighter integration with Claude; no local control.
Sourcegraph (on-premise or SaaS)
Mature code intelligence platform with advanced search, refactoring, and monitoring. Requires more infrastructure and operational overhead than Codanna but offers greater scalability and commercial support.
Build on codanna with DEV.co software developers
Evaluate Codanna for your codebase. Test semantic search, relationship tracking, and MCP integration on a sample project. Review Apache 2.0 licensing and Windows support status for your environment.
Talk to DEV.coRelated open-source tools
Surfaced by semantic similarity across the DEV.co open-source index.
Related on DEV.co
Explore the category and the services that help you build with it.
codanna FAQ
Does Codanna send code to the cloud?
Which languages are supported?
Can I use Codanna with my IDE?
What happens if the index is outdated?
Software development & web development with DEV.co
Need help beyond evaluating codanna? DEV.co is a software development agency offering software development services and web development for teams of every size. Our software developers and web developers build custom software, web applications, APIs, and rag frameworks integrations — and maintain them long-term.
Ready to give your AI agents smarter code understanding?
Evaluate Codanna for your codebase. Test semantic search, relationship tracking, and MCP integration on a sample project. Review Apache 2.0 licensing and Windows support status for your environment.