rag-rat
What a repository knows about itself. rag-rat is a local repo-intelligence index and MCP server
for coding agents. It keeps source files read-only, writes only its own SQLite database, and answers
with provenance on every result — current source, the code graph, git/GitHub history, and durable,
source-anchored repo memories that persist across sessions and agents.
Every coding harness already has grep and file reads. rag-rat adds the layer they do not provide:
source-anchored rationale. It connects the code an agent is about to touch to its callers, callees,
tests, git/GitHub history, prior decisions, invariants, risks, and duplicate-code signals — and
labels every result with confidence and coverage, so an agent can judge it instead of trusting it.
sequenceDiagram
participant Repo as Repository
participant Engine as rag-rat engine
participant Agent as Coding agent
Repo->>Engine: Source · git/GitHub · repo memories
Engine->>Engine: Index → graph → (opt) SCIP oracle → reconcile
Agent->>Engine: where / why / who-calls / impact?
Engine-->>Agent: source + call paths + papertrail + memories (with provenance)
Agent->>Engine: record a finding
Engine->>Repo: persist a source-anchored repo memory
Why
- Provenance, not guesses. Every result carries a confidence label, coverage warnings, and the raw evidence — so a partial index or an ambiguous edge reads as exactly that.
- Repo memories. Typed, source-anchored notes (
Invariant,Decision,Risk, …) that survive refactors and surface automatically during future queries — the signal grep can't give you. They are not assistant memory: they are versioned, local, source-anchored facts about this repository that any future agent retrieves with evidence. - A real code graph. tree-sitter callers/callees/imports across Rust, TypeScript/TSX, Kotlin,
C/C++, and Python — with an optional compiler-grade SCIP oracle that upgrades
edges to
Compilerconfidence and ranks the load-bearing symbols. - History as evidence. Git history, lazy chunk blame, and cached GitHub issue/PR/review rationale, all queryable.
- Rides your existing grep. A PreToolUse hook injects the memories and symbols behind whatever you just searched for.
- Flags clones as you write them. A PreToolUse hook on Write/Edit/MultiEdit fingerprints the functions you're writing and warns when they're exact or near-duplicates of code already in the repo — so an agent reuses instead of re-implementing. Read-only, and a silent no-op when the index isn't ready, so it never blocks a write.
Install
From a checkout:
Add --no-default-features for a smaller hash-only build without real embeddings. SQLite is bundled
(compiled in via rusqlite), so there is no system-library prerequisite — see
Platform support for the per-OS C-toolchain note.
Quickstart
From the repository you want to index:
init scans the repo, prompts for languages and path bindings, writes rag-rat.toml, indexes,
offers to install the local embedding model, and can register the MCP server and git hooks. Preview
without writing anything with rag-rat init --dry-run; --yes runs the non-interactive defaults.
Manual setup and every config knob live in docs/config.md. For a large repo where
the default local embedder is too slow, see Embedding backends.
Connect it to your agent (MCP)
The MCP server is STDIO — the client launches rag-rat as a child process. rag-rat init is the
recommended path: it registers the server per project (claude mcp add --scope project /
codex mcp add), so each repo gets its own index.
To wire it up by hand, register a project-scoped server that runs in the repo directory:
or a project .mcp.json / equivalent:
Don't pin a single global server to one repo's config. A user-scoped server with a hardcoded
--config /some/repo/rag-rat.tomlserves that repo's index and memories everywhere — so browsing a different codebase loads the wrong context. Register the server per project and let it resolverag-rat.tomlfrom the repo it runs in. (--config <path>still exists for the rare case you need to point at a specific profile.)
Pass rag-rat mcp --json if your client must parse tool text as JSON (results are
TOON by default). Full tool schemas: docs/mcp-tools.md.
Try it
Right after rag-rat init the code graph, symbols, git history, semantic search, and clone
detection are ready — these answer on the first query. Repo memories start empty: they accrue as
agents record findings with memory_create and then surface automatically in later answers. (GitHub
issue/PR rationale needs a rag-rat github sync.)
Ask your MCP client:
- "Run
impact_surfaceon the function I'm about to edit — its callers, callees, tests, and recent commits." - "Where is config reload handled?" — hybrid
semantic_searchover source and docs. - "What are the most load-bearing symbols in this repo?" —
important_symbols. - "Does this helper duplicate anything already in the codebase?" —
find_clones(and the write-time hook warns as you write it). - "Record an invariant on
parse_config: reload must not allocate after the scheduler starts." —memory_createwrites your first repo memory; it then rides along in futureimpact_surface/symbol_lookupresults.
Or from the CLI:
The agent loop
The point isn't the tool catalog — it's the loop an agent runs around an edit, so it changes code with the callers, tests, rationale, and prior art in front of it instead of guessing:
- Before editing a symbol, ask
impact_surface. One call returns the current source anchor, callers and callees, related tests, git/GitHub rationale, the repo memories bound to that symbol / path / call-path, and confidence + coverage warnings. - Read the blast radius, then edit. The invariant a previous agent recorded, the caller three hops away, the test that pins the behavior — all surfaced before the change, not discovered after.
- The clone hook catches duplication at write time. If the new function reimplements code that already exists, the Write/Edit hook says so, with the existing symbol to reuse.
- Record what you learned. When the edit reveals a durable invariant, decision, or footgun,
memory_createstores it as a source-anchored repo memory — so the next agent (or the next session) gets it in one call instead of re-deriving it.
A trimmed impact_surface answer (TOON — the default output; abbreviated here) — every field is
evidence, not prose:
query:
ref: "crates/config/src/config.rs::parse_config"
resolution: syntactic
direct_semantic_callers[12]:
- from_symbol: "crates/runtime/src/boot.rs::start"
edge_kind: calls_name
confidence: syntactic
callsite:
path: "crates/runtime/src/boot.rs"
line: 88
importance:
label: local structural load
score: 6.8
bucket: high
tests_touching_symbol_path[4]:
- path: "crates/config/src/config_tests.rs"
reason: test_mentions_symbol_or_path
recent_commits_touching_symbol_path[1]:
- evidence[1]: "a1b2c3d touched crates/config/src/config.rs: fix reload race during startup (#141)"
repo_memories:
direct[2]:
- kind: Invariant
title: "Config reload must not allocate after the scheduler starts"
confidence: high
anchor_status: current
binding_kind: symbol
- kind: Decision
title: "TOML over JSON5 for the config surface (#88)"
anchor_status: current
binding_kind: path
completeness_and_caveats:
exact_graph_callers: 12
memory_status:
active: 2
stale: 0
caveats[1]: "Graph evidence is tree-sitter/syntactic, not compiler-grade name resolution."
And the write-time clone warning an agent sees before it duplicates logic — verbatim hook output:
▶ rag-rat clone check — code you're writing duplicates existing functions:
• `normalize_path_for_lookup` (line 42) is ~91% similar to crates/index/src/paths.rs::canonicalize_lookup_path
Prefer reusing the existing function(s) over duplicating — impact_surface / symbol_lookup to inspect them.
The tools
The highest-leverage ones (full catalog + JSON schemas in docs/mcp-tools.md):
impact_surface— the coding preflight from the loop above: callers, callees, tests, git history, GitHub papertrail, and repo memories for a symbol in one call.repo_memoriesdefaults to a compact, scannable per-memory header (kind, title, confidence, anchor status, and where it's bound); passfull_memories: true(or usememory_for_symbol|path|call_path) for full bodies + bindings.semantic_search— hybrid BM25 + vector recall over source/docs, validated against current source. Every hit reportsretrieval_mode;explain=truebreaks down the score.symbol_lookup— exact/fuzzy symbol resolution; cfg/overload duplicates grouped as one logical symbol.find_callers/trace_callees— reverse/forward graph traversal (low-signal std/macro noise filtered by default).important_symbols— load-bearing symbols by (SCIP-aware) PageRank; seedocs/oracle.md.repo_brief/repo_clusters— orientation: spine / churn / god-modules / ownership clusters.find_clones/clones_for_symbol— exact + near-miss duplicate functions ranked by refactor ROI; the candidate graph is precomputed in the background so it scales to large repos.read_chunk— current text for a chunk with anchor validation.- Git/GitHub:
commit_search,git_history_for_path|symbol,git_blame_chunk,papertrail_for_*,rationale_search. - Memories:
memory_create,memory_update,memory_search,memory_for_symbol|path|call_path,memory_validate,memory_mark_obsolete.
Repo memories
Repo memories are first-class local evidence — not chat memory, not cloud personalization. They
are versioned, local, source-anchored facts about this repository. Each is typed
(Invariant, Decision, RejectedAlternative, Risk, BugPattern, PerformanceNote, …) and
source-anchored: bound to a logical symbol, concrete symbol, chunk, path+span, graph edge,
call-path, commit, or GitHub ref. rag-rat tracks each anchor as current, relocated, stale,
gone, or unverified, and surfaces matching memories through the memory_* tools and inline in
read_chunk, symbol_lookup, find_callers, trace_callees, and impact_surface. They're how
hard-won context reaches the next agent in one call instead of evaporating.
Compiler-grade resolution & ranking
The graph is heuristic by default. The opt-in SCIP oracle (rag-rat oracle run) upgrades edges
to a Compiler tier from a real language tool, recovers calls tree-sitter missed, flags external
edges, and makes important_symbols surface the genuine god-modules. For C/C++ the scip-clang
oracle distinguishes declarations from definitions and sharpens call/type edges in macro-heavy or
multi-target code — the difference between usable and noisy graphs on firmware, kernels, drivers, and
SDKs. Turn on [oracle] auto_run and the MCP server keeps it fresh on its own (throttled,
watcher-safe). Full details: docs/oracle.md.
Freshness
rag-rat mcp runs a background file watcher (on by default; [watch] enabled = false or
RAG_RAT_NO_WATCH=1 to disable), so graph/symbol queries reflect uncommitted edits without a commit.
Indexed rows are git-context-aware: clean files are stored by commit_sha, dirty/untracked files in
a worktree overlay, so one database reuses rows across branch switches while reflecting local edits.
Optional git hooks (rag-rat hooks install) keep the index current on checkout/merge/rewrite/commit.
read_chunk and search validate hits against current source and heal stale entries before returning.
One watcher per worktree and one writer at a time are enforced with file locks (unreliable on
NFS / WSL2 /mnt mounts).
Output format
The CLI and MCP results default to TOON (Token-Oriented Object Notation) — a token-efficient
encoding that renders uniform rows as a dense [N]{cols}: table (~30% smaller than compact JSON on
those payloads, never larger in practice). Pass --json (CLI, either position) or launch
rag-rat mcp --json (MCP) when a JSON parser must read the output.
Embedding backends
The default local embedder (FastEmbed) needs no setup, but a large repo or a stronger model is worth
offloading. rag-rat speaks the OpenAI-compatible /v1/embeddings API, so a [llm.embedding.remote]
block can serve embeddings from Ollama, vLLM, or michaelfeil/infinity — one client, one place to
audit and secure. Two modes:
- Connect to a server you already run (set
endpoint). - Ephemeral — let the bundled cookbook provision a GPU worker (Modal / RunPod) just for the
backfill and tear it down afterward (set
cookbook); pick the backend and GPU class in config.
The init flow warns when a short-context model would truncate long code chunks and steers you to a
long-context code embedder, and rag-rat auto-tunes the client concurrency against the chosen
backend so the sweep finds its throughput knee. Setup and every knob: docs/config.md.
Retrieval quality
Search quality is measurable, not guesswork. rag-rat ships a commit-replay evaluation harness
(rag-rat eval --replay): each recent commit becomes a case — its message is the query, the files it
touched are the gold set — and search is scored on how well it recovers them. It reports recall@3
(did the right chunk land in the first three reads?), recall@10, and MRR@10, and CI tracks the trend
on Bencher on main so a regression is caught before it
ships.
Reach for it when comparing embedding models, changing chunking, enabling int8 vector storage
(smaller on disk), or tuning a remote backend — you can prove the change didn't cost recall instead
of hoping. (rag-rat eval requires a --features eval build; it is absent from the released binary.)
Benchmarks
The headline workload is indexing the whole Linux kernel (v7.0, ~63k C/H files, 11.2M graph edges).
Full numbers — wall-clock, throughput, peak RSS, on-disk size, unresolved-edge taxonomy — are in
docs/benchmarks.md. Performance is tracked per-push and gated per-PR; the live
history is at bencher.dev/perf/rag-rat/plots (wiring:
docs/bencher.md).
Security
The MCP server exposes read-only source tools. It never executes shell commands or writes your source
files. It writes only the configured SQLite index — during indexing, migration, maintenance,
reconciliation, repo-memory operations, and automatic stale-index healing. GitHub sync is explicit
and uses gh api; normal query tools read only the local cache.
Local vs remote embedding
With the default local embedder, nothing leaves the machine — indexing and querying are entirely
local. Configuring a [llm.embedding.remote] backend is what sends text off the box, in two places:
the chunk text selected at index time, and the query text of each semantic search (a search
embeds your query to compare it against the indexed vectors). A CONNECT backend embeds both against
the configured endpoint; an ephemeral backend embeds queries against the local query_endpoint.
What the endpoint is decides how much that matters:
- Your own server (self-hosted Ollama / vLLM / infinity) — the text stays in infrastructure you control.
- Ephemeral Modal / RunPod workers (the cookbook path) are ephemeral compute providers running your open-source embedder, not data services that train on inputs. Both are SOC 2 Type II, encrypt in transit and at rest, isolate tenants, and tear the box and its storage down after the backfill — a data-processor relationship, reasonable for proprietary code the same way a cloud VM is.
- A third-party embedding API you don't control is the one to actually read the terms on (retention, training on inputs).
Sensible hygiene regardless of backend: exclude secrets, generated files, and vendor trees from the
indexed targets so they're never chunked or embedded, and keep secrets out of query text. Details:
docs/config.md.
Platform support
rag-rat builds and tests on Linux, macOS, and Windows. Linux is covered on every PR and on every
push to main; macOS and Windows are exercised on release, so cargo install rag-rat builds and
links on all three.
SQLite is bundled (compiled from source via rusqlite), so there's no system-library prerequisite,
but each platform needs a C toolchain: Linux ships one; on macOS install the Xcode Command Line
Tools (xcode-select --install); on Windows install the Visual Studio Build Tools with the C++
workload (MSVC). Requires Rust 1.95+ (the bundled SQLite build uses the cfg_select! macro,
stabilized in 1.95).
A few maintenance conveniences are Unix- or Linux-only by design and degrade quietly elsewhere — no feature of the index, query, or MCP surface is affected:
- Hot-upgrade of a running MCP server (the
SIGUSR1in-place re-exec) is Unix-only. On Windows, restartrag-rat mcpto pick up a new binary. - Fleet auto-upgrade (signalling other running servers when a new binary lands) is Linux-only —
it walks
/proc— and is a no-op elsewhere. - The grep-augmentation hook uses a warm Unix-socket listener (with per-session dedupe) on Linux and macOS; on Windows it falls back to a per-call read-only query straight against the index, which works the same but without cross-call dedupe.
Commands
|||
| |
Releasing & license
Releases are automated by release-plz (the three crates ship in lockstep;
see docs/releasing.md). rag-rat is MIT-licensed — see LICENSE.
Prior art
rag-rat's clone-detection design is inspired by SourcererCC's scalable token-bag candidate generation, NiCad's normalized near-miss clone-detection framing, GumTree's move-aware AST differencing, and anti-unification / least-general generalization for template extraction. Planned fragment-level mining and copy-paste bug heuristics are inspired by CP-Miner.