Compare Code Intelligence Tools for AI Agents

Real benchmarks on the same hardware. Not marketing claims — live measurements on the same codebases including torvalds/linux at 74,619 files. All competitor data source-verified — every claim below was verified by indexing and reading the competitor's source code directly through Synapse MCP.

37.5×
Faster Discovery
vs Graphify on Linux kernel
5.9×
Less RAM
877 MB vs 5.21 GB peak
400×
Faster Queries
148 ms vs 59.3s cold query
9.7–15.2×
Fewer Tokens
vs Graphify on same queries

Feature Comparison

FeatureSynapse MCPCodeGraphcode-graph-mcpGraphifySourcegraphGreptileAugment Code
ArchitectureElixir BEAM · ETS in-memory · persistent daemon · multi-repoTypeScript host + Rust kernel (napi-rs) · per-projectPure Rust single binary · per-projectPython · NetworkX · file-based graph.json · no daemonGo · Zoekt index · SaaS/self-hostedVector RAG · cloud SaaSCloud RAG · proprietary index
StorageETS tables + SQLCipher SQLite · persistent daemonSQLite (FTS5)SQLite (FTS5 + sqlite-vec)Single graph.json fileZoekt shards on diskProprietary vector DBProprietary index
Query SpeedSub-ms ETS reads · 148ms on Linux kernel~50ms (SQLite FTS5)~30ms (SQLite RRF fusion)59.3s cold start (1.6 GB JSON)~100ms (Zoekt)~500ms~200ms
SearchNative BM25F + random projection (in-process, 10-50µs/chunk) + symbol + regexFTS5 + symbol. No embeddings.BM25 + vector RRF (external Candle model, 419 deps)BFS neighbourhood + FTS5Zoekt indexVector similarityProprietary
Languages50+ (source-verified: 90+ extensions, 691 LOC registry)20 full + 10 extended + 17 web frameworks (incl. Elixir? No)20 (source-verified: src/parser/languages.rs)36+ (tree-sitter)30+20+30+
Write SafetyEXCLUSIVE Shadow Graph — validates before disk writeRead-onlyRead-onlyRead-onlySearch onlySearch onlySearch only
Crash ResolutionEXCLUSIVE resolve_stack maps traces to AST chunksNoneNoneNoneNoneNoneNone
Dead Code DecisionsGraph-aware: safe_to_delete, confidence, caller splitting (prod vs test), agent instructionsNot documentedSQL-based w/ per-language exclusions (intentionally unlisted from MCP)Lists zero-caller functionsNoneNoneNone
Change / Diff ReviewEXCLUSIVE Pre-change blast radius + risk scores + test mappingNoneNonePost-hoc git diffReference searchSearch onlySearch only
Token CompressionSmartCrusher: 30-60% reductionNoneNoneNoneNoneNoneNone
HTTP Route TracingVia trace_behaviour (Plug/middleware chain, runtime-verified)UNIQUE 17 web frameworks with heuristic URL→controller→template edgesExpress, Flask, Go, axum via route_pathNoneNoneNoneNone
Cross-Language BridgingNative SwiftChunker + ObjectiveCChunker (AST-depth, no synthetic bridge edges)UNIQUE Swift↔ObjC, React Native bridge, Expo, FabricSwift parsed (tree-sitter), no bridgingNoneNoneNoneNone
Git WorktreesVWO: <50ms delta indexingPer-project init onlyPer-project onlyGRAPHIFY_OUT per worktreeNoneNoneNone
Multi-RepoEXCLUSIVE Registry, worktrees, lifecycle statesPer-project onlyPer-project onlyPer-projectPer-instancePer-orgPer-org
Query During IndexingImmediate (2s from registration)Partial (during resolution)Partial (Merkle re-check)Blocked until build completePost-indexPost-ingestPost-index
Cross-Session LearningKnowledge cache: fingerprinting, hit tracking, learningNoneNoneFile-based reflect.py (offline)NoneNoneNone
BDD/Schema IngestionGherkin · SQL DDL · Protobuf · GraphQL · HCLCode onlyCode onlyMarkdown onlyCode onlyCode onlyCode only
IDE Auto-ConfigRust CLI: 20+ IDEs, 16 rule targetsVia prompt hook config.claude/plugin + hooks.jsonPython: 17 assistantsManualManualManual
Elixir/BEAM SupportNative Macro.prewalk + alias resolutionNot listedNot listedNo alias resolverBasicNoneNone
Privacy100% local · zero cloud · zero telemetry100% local100% localOSS: local · SaaS: cloud uploadSaaS/self-hostedCloud onlyCloud only
PricingFree (full product) · Pro $19/moMIT — freeMIT — freeFree tier · Pro $10/mo · Teams $20/seatFree tier · Enterprise customFree tier · Pro $24/moFree tier · Teams $30/seat
Linux Kernel RAM877 MBN/AN/A5.21 GBN/AN/AN/A
Cold Query (Linux)148 msN/AN/A59.3 secondsN/AN/AN/A

Synapse vs Graphify benchmarks from live physical tests on same Apple Silicon hardware. Linux kernel: 74,619 code files. CodeGraph and code-graph-mcp feature data source-verified by indexing their source repositories through Synapse MCP (September 2026). Other products: publicly documented capabilities. EXCLUSIVE = no competitor offers this. UNIQUE = only this competitor offers this.

Linux Kernel Stress Test — Synapse vs Graphify

The only two tools tested on torvalds/linux (74,619 files) on the same hardware:

MetricSynapse MCPGraphifyWinner
Filesystem discovery2.0s1m 15sSynapse — 37.5×
Peak RAM877 MB5.21–5.75 GBSynapse — 5.9× lower
Query during indexingAvailableBlockedSynapse
Cold query (EXPORT_SYMBOL)148.1 ms59.3 secondsSynapse — 400×
CPU scheduling24 BEAM schedulersSingle core 100% for 15+ minSynapse
OOB scalabilityZero configCrashed: needed GRAPHIFY_MAX_GRAPH_BYTES=4GBSynapse

Full benchmark data and methodology at synapse-mcp.dev/blog/synapse-vs-graphify-live-benchmark.html.

Deep-Dive: CodeGraph & code-graph-mcp — Source-Verified Comparison

In September 2026, we source-verified every claim below by indexing all three codebases through Synapse MCP and reading the actual source code. No marketing claims. No assumptions. Verified implementations.

Research methodology: All three MCP code-graph tools were indexed by Synapse MCP and their source code read directly. Synapse MCP: 12,775 chunks / 1,423 files / Elixir umbrella (3 OTP apps). CodeGraph (colbymchenry): 15,789 chunks / 356 files / TypeScript host + Rust kernel. code-graph-mcp (sdsrss): 1,484 chunks / 229 files / Pure Rust single binary.

How They Compare on Architecture

CodeGraph (colbymchenry) is the most feature-complete read-only competitor. It runs as a TypeScript/Node host with a compiled Rust native kernel via napi-rs, stores its graph in SQLite with FTS5, and has the broadest language coverage of any read-only tool: 20 full languages, 10 extended, plus 17 web frameworks (Django, Rails, Express, Spring, Go Fiber, FastAPI, axum). Its unique capability is HTTP route-chain tracing — following a URL through routing layers into the handler for 17 frameworks using synthesised heuristic call-graph edges. It also supports cross-language bridging for Swift↔ObjC and React Native JS↔Native boundaries. What it lacks: no vector/semantic search, no write safety, no multi-repo support, no test mapping, no debugging.

code-graph-mcp (sdsrss) is a pure Rust single binary (165 source files) with the most sophisticated indexer of all three: a BLAKE3 Merkle tree for incremental change detection, notify-based OS file watching, hybrid BM25 + vector search via Reciprocal Rank Fusion with an embedded Candle (Rust ML) model, and production-grade SQLite (WAL mode, schema versioning, FTS5, sqlite-vec). Its dead code detection is a 2,551 LOC SQL query module with per-language constructor exclusions — intentionally unlisted from the MCP tool surface (confirmed by a test assertion). It is also read-only and per-project only.

Verdict Summary — Synapse Wins Where It Matters

Search Quality
Synapse MCP
Native BM25F + random projection in-process at 10-50 µs/chunk, live IDF corpus, zero external dependency. code-graph-mcp's Candle model requires external binary + 419 dependencies requiring daily cargo audit.
Dead Code Detection
Synapse MCP
Graph-aware: production vs test caller splitting, confidence tiers, deletion-risk reasoning, agent instructions, next_tool_calls. code-graph-mcp has better per-language SQL exclusions; Synapse wins on analytical depth.
Language Breadth
Synapse MCP
90+ extensions / 50+ languages in @detection_hash (source-verified, 691 LOC). Covers languages no competitor does: Clojure, Haskell, Solidity, COBOL, Fortran, Zig, Assembly, Erlang, F#, OCaml, Verilog, SQL DDL, Protobuf, GraphQL, HCL, Gherkin.
Web Framework Routing
CodeGraph (narrow use case)
CodeGraph synthesises URL→controller edges for 17 frameworks in one call. Synapse's trace_behaviour goes deeper (full Plug/middleware/callback chain with runtime verification) but requires knowing your entry point.
Write Safety
Synapse MCP EXCLUSIVE
No competitor offers any write capability. Synapse's Shadow Graph simulates edits in memory, validates with linter, atomically commits only on success. Neither CodeGraph nor code-graph-mcp can prevent agents from breaking production.
Change Review
Synapse MCP EXCLUSIVE
Pre-merge blast radius, broken contract detection, test mapping — before the PR lands. No competitor offers diff analysis, impact scoring, or ranked risk.
Multi-Repo / Worktrees
Synapse MCP EXCLUSIVE
Registered multi-repo with lifecycle states, Git worktree detection, delta indexing. No competitor supports more than one project per process.
Engineering Rigour
Synapse / code-graph-mcp
code-graph-mcp: 100+ versioned changelog entries, BLAKE3-pinned vendored C sources, compile-time budget guard. Synapse: mix deps.audit on every push (stricter than daily cron), multi-platform smoke tests, BDD e2e, signed release manifest with launcher version gate.

The Honest Summary

code-graph-mcp is a serious competitor with excellent engineering discipline: BLAKE3-pinned vendored C sources, compile-time instruction budget guards, a rich changelog. Its dead code SQL and Merkle-based incremental indexer are well-designed. Where it falls short: read-only, per-project only, 20 languages (no Elixir, Erlang, Haskell, OCaml, Scala, Solidity, COBOL, Zig, Gherkin, HCL, Protobuf, GraphQL, SQL DDL), no agent write safety. Its cargo audit runs on a daily cron; Synapse's mix deps.audit runs on every push.

CodeGraph has two capabilities no competitor matches: HTTP route-chain tracing for 17 frameworks and React Native / Expo / Fabric bridge-aware graph traversal. Where it falls short: no embeddings, no semantic search, read-only, per-project only, no write safety, no change review, no persistent memory.

Synapse wins on: language breadth (50+ languages, 90+ extensions), framework detection (30+ indicators), native Swift/ObjC indexing at AST depth, embedding quality (native BM25F, zero external dependency, 10-50 µs), zero-latency file event indexing, behaviour tracing (full Plug/middleware chain with runtime verification), dead code analytical depth, and the entire write-safety / change-review / multi-repo / worktree capability cluster that no competitor is building toward.

Synapse MCP — Free, Forever

Full AST indexing, 50+ languages, safe writes, SmartCrusher compression, Git worktree overlay, unlimited repos. No credit card. No cloud.

Download Now — for FREE →