codegraph vs code-review-graph
Two of the top projects in MCP servers for developers, side by side: who leads on each score part and who each one suits, then setup, license, activity and what each review found.
codegraph
Local code knowledge graph that coding agents query over MCP
code-review-graph
Local Tree-sitter code graph that gives coding assistants review context over MCP
Verdict
codegraph ranks higher, 95 to 92. Of the 4 score parts we measure for both, codegraph leads on 2, code-review-graph on 1 and 1 is even.
- Adoptioncodegraph leads by 8 points, 98 to 90.
- FreshnessEven at 100.
- Maintenancecode-review-graph leads by 2 points, 93 to 91: close to even.
- Agent-readycodegraph leads by 15 points, 85 to 70: a wide gap.
Pick codegraph if you are
Developers who want faster, cheaper code navigation for AI coding agents
- Auto-configures nine agents with one `codegraph install` command
- Index stays current by watching files; no manual re-indexing
Pick code-review-graph if you are
Developers who want coding assistants to read fewer files when reviewing changes
- One install command detects and configures about 16 AI coding tools, with a clean uninstall
- Incremental updates re-parse only SHA-256-changed files; two-file edit on django took about 2.5 s
| What we compare | codegraph | code-review-graph |
|---|---|---|
| Score parts, out of 100 | ||
| Adoption | 98, widely used | 90, widely used |
| Momentum | Not scored for this kind | Not scored for this kind |
| Freshness | 100, active | 100, active |
| Maintenance | 91, healthy | 93, healthy |
| Agent-ready | 85, ready | 70, partly |
| Facts from GitHub and the README | ||
| Stars | 77.4k | 32.7k |
| License | MIT (permissive) | MIT (permissive) |
| Last commit | Nov 2026 | Nov 2026 |
| Last release | Oct 2026 | Nov 2026 |
| Language | TypeScript | Python |
| Docker | No | Yes |
| GPU | Not needed | Not needed |
| arm64 or Apple Silicon | Mentioned | Not stated |
codegraph
CodeGraph indexes a project into a local `.codegraph/` graph of symbols, call edges and dependencies, then exposes it to coding agents through an MCP server. Agents can fetch relevant source, call paths and change impact in one call instead of grepping and reading files. `codegraph install` wires it into Claude Code, Cursor, Codex, opencode, Gemini, Kiro, Copilot and others; the index auto-syncs on file changes.
Who it is for: Developers who want faster, cheaper code navigation for AI coding agents
Strengths
- Auto-configures nine agents with one `codegraph install` command
- Index stays current by watching files; no manual re-indexing
- Covers about 35 languages with cross-file resolution in one graph
- Benchmarks published with method: 7 repos, median of 4 runs, CLI blocked in both arms
Weaknesses
- Benchmark figures come from the author and use one model, Claude Opus 4.8
- README says responses leave about 80% more context resident in long sessions
- Must run `codegraph init` in every project to build its graph
- Hosted CodeGraph platform is a waitlist product, separate from this tool
- no GPU
- Models: Claude Code, Cursor, Codex CLI, opencode, Hermes Agent
code-review-graph
Parses a repository with Tree-sitter into a SQLite graph of functions, classes, imports, calls and tests, then serves blast-radius and risk-scored context over MCP so the assistant reads only affected files. The installer configures about 16 tools (Claude Code, Cursor, Codex, Windsurf, Zed and others) and updates incrementally via hooks, git hooks or watch mode. A GitHub Action posts risk-scored PR comments.
Who it is for: Developers who want coding assistants to read fewer files when reviewing changes
Strengths
- One install command detects and configures about 16 AI coding tools, with a clean uninstall
- Incremental updates re-parse only SHA-256-changed files; two-file edit on django took about 2.5 s
- Covers 30+ languages plus notebooks, and custom grammars via languages.toml
- Runs locally; the GitHub Action builds the graph on the CI runner without sending code out
Weaknesses
- Impact recall of 1.0 is circular, since ground truth comes from the same graph
- Against a grep baseline, median token reduction is 6.65x, not the 63x headline
- Graph context can cost more tokens than a plain read for small single-file edits
- Cold build of a roughly 3,000-file repository took about 40 seconds
- no GPU
- Docker
- Needs Python 3.10+, Tree-sitter, SQLite
- Models: all-MiniLM-L6-v2 (local embeddings)
More in MCP servers for developers
Building with these?
We help teams pick an open-source AI stack and run it in production. Client work never changes a rank.