repowise
A read-only codebase-intelligence server that answers questions about a repository's code, history, architecture, and health from a prebuilt index, so an agent can query instead of grepping and reading files.
Answer questions in one call —
get_answerreturns an evidence-grounded, cited response with a confidence level and candidate files to open next.Get oriented in an unfamiliar repo —
get_overviewgives packages, entry points, riskiest files, and top actions.Triage files, modules, or symbols —
get_contextreturns summaries, signatures with line numbers, hotspots, callers/callees, ownership, decisions, and doc drift; batch many targets at once.Search the codebase —
search_codebasedoes hybrid (full-text, semantic, symbol, path) search and can resolve identifiers to exact symbol IDs or list files to read.Read exact source —
get_symbolreturns line-verified symbol bodies, optionally with transitive callee bodies or a line range.Assess change danger —
get_riskfuses git signals (hotspots, churn, ownership, bug history, co-change partners) and graph reach; PR mode returns a directive of what may break, missing companion files, and tests to run.Review a commit, range, or uncommitted work —
get_change_riskgives a verdict, health delta, risk percentile, diff shape, and impacted tests.Find code health issues and fixes —
get_healthreports 1–10 scores, findings across defect/maintainability/performance, a ranked fix-first queue, refactoring plans, coverage, trends, and doc drift.Locate dead code —
get_dead_codelists unreachable files, unused exports, and zombie packages tiered by confidence, with deletion-ready flags.Recover design rationale —
get_whysurfaces decision records and evidence commits behind the code, falling back to git archaeology.
All tools are read-only and local, with no LLM calls or external network access.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@repowise show the dependency graph for the auth module"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Set it up in one line
Open Claude Code, Codex, Cursor or any MCP-capable agent in your repository and paste:
Read https://docs.repowise.dev/setup.md and set up Repowise in this repository.The agent installs Repowise, indexes the repo with no API key, connects itself to the index, and asks you before anything costs money. To do it by hand:
uv tool install repowise # or: pipx install repowise / pip install repowise
cd /path/to/your/repo
repowise init --yes --no-prose # graph, git, health, dead code, docs. No key, no spend.
# answers and docs: run repowise generate (one command, no re-index)
repowise serve # dashboard + MCP server (needs Node 20+; first run downloads the UI)serve opens the dashboard in your browser. After a first index it starts with a short
guided tour, and Setup is where you connect agents, turn on sync and add coverage or a model.
Restart Claude Code or run /mcp, then ask "What breaks if I change
src/auth.py?" or "Use Repowise get_overview to summarize this repository."
init wires Claude Code for this repo and adds the global repowise MCP entry
when there is none; it never moves one that serves another repo. Full setup for every agent →
Related MCP server: CodeGraph
Your agent stops guessing
Without an index, an agent rediscovers your codebase on every task: grep, open a file, open three more, grep again. Repowise answers those questions ahead of time, so the agent asks once and gets on with the work.
In a measured run on django/django, an agent with Repowise made 3.8 tool calls
instead of 7.2, opened 3.0 files instead of 7.2 and wrote 31.6% less, with
answers just as good. That is the largest saving of the five context tools we
tested. On a small local model the output cut reaches 47.9%.
How it was measured →
Your agent gets ten task-shaped MCP tools. Ask about several files at once and get callers, risk, tests, owners and the reasons behind the code in one answer. The tool list ↓
Up to 99% less command noise.
repowise distilltrims test, install and git output before the agent reads it and keeps every error line. Distill →Context before the agent asks. Optional hooks warn it off fragile files and stale paths, and Repowise keeps your
CLAUDE.mdandAGENTS.mdcurrent from the index.See your own savings.
repowise savedand the dashboard's Usage and savings page count the tokens your agent no longer has to read.
What it does
One index of your code, its history and its decisions, used three ways: to understand the code, to change it safely, and to keep improving it.
Understand the code
You ask | Repowise gives you |
How does checkout work in this repo? | A cited answer built from the call graph and the generated docs, in one call. Search and answers |
What calls this function, and what does it call? | A call graph across 28 parsed languages, every edge stamped with how it was resolved and how far to trust it, plus traced execution flows from each entry point. The graph |
Can I get docs for this codebase? | A wiki for every module and file, rendered from the code's structure with no key, or written by a model when you choose. It updates incrementally after each commit. Docs |
Which of our docs are wrong? | Markdown checked against the tree: every reference to a file or symbol the code no longer has, with the line to edit. Doc drift |
Why is it built this way? | Decisions mined from ADRs, |
Who knows this code? | Owners, bus factor, fading ownership when the main author goes quiet, and suggested reviewers. Ownership |
Can I see the architecture? | An explorable dependency map, C4 views, and a Structurizr export, no model involved. Dashboard |
Change it safely
You ask | Repowise gives you |
What breaks if I change this? | Symbol-level blast radius: the callers of what you changed, the files that historically change with it but are missing from your diff, and the tests that reach it. Change risk |
How risky is this PR? | Where the change ranks against your repository's own recent commits, with the reasons, as a directive your agent can act on. Change risk |
Which tests should run? | The tests a diff actually exercises, from a coverage report if you have one and from the call graph if you do not. Test intelligence |
Did this PR add untested lines? | Patch coverage, branch coverage on changed lines and path-scoped gates in CI, on GitHub, GitLab or any runner. CI gates |
Will this API change break another repo? | HTTP, gRPC, topic and OpenAPI contracts matched across repositories, with a breaking-change guard and the consumer files it affects. Workspaces |
Is anyone else editing these files? | Other open branches touching the same files or their co-change partners, each with the reason it is listed. Branch overlap |
Did we just commit a secret? | Keys, tokens and risky calls found in the working tree and in full git history, with a pre-commit check and a CI gate. Security signals |
Improve it continuously
You ask | Repowise gives you |
What should we fix first? | A ranked queue weighing impact against effort, using churn, fan-in, coverage and bug history. Fix first |
Where is the debt? | A 1 to 10 score for every file from 53 deterministic detectors, split into defect risk, maintainability and performance, validated against real bug history. Code health |
Why is this slow? | N+1 queries, I/O in loops, blocking calls inside async code and quadratic loops, traced across function and file boundaries. Performance |
How do I break this up safely? | Concrete refactoring plans: Extract Method, Extract Class, Move Method, Split File, Break Cycle, with the exact symbols that move and what moves with them. Ready to hand to an agent. Refactoring |
What can we delete? | Unreachable files, unused exports and unused packages, each with a confidence tier and the evidence behind it. Dead code |
Where do bugs keep landing? | Bug-fix commits traced to files and symbols, and a warning when your agent edits a bug magnet. Bug history |
Are our tests testing anything? | Tests with no assertions, tests that only check their own mocks, and untested hotspots. Test-quality smells |
Test coverage without running coverage. Most repositories never produce a coverage report. Repowise answers "is this tested, and by what" from the import and call graph, and labels every answer measured or inferred.
It checks its own health score on your repository. After each index it reports how many of the lowest-scoring files actually had bug fixes in your recent history, so a bad score on your codebase is visible to you.
It learns from your agent sessions. With transcript capture on, the corrections you keep repeating ("use the shared HTTP client") become tracked decisions it feeds back to the agent later. Transcripts never leave your machine.
It writes your
CLAUDE.mdandAGENTS.mdfrom the real index and keeps them current, so even an agent with no MCP support starts informed.It shrinks command output before your agent reads it.
repowise distill pytestkeeps every failure and drops the noise, and nothing is lost: anexpandcommand restores any cut.New git worktrees start indexed. A linked worktree seeds its index from the base checkout instead of starting cold.
It knows which external systems you depend on. Package manifests across PyPI, npm, Cargo, Go, NuGet, Maven and CMake feed a map of the services and libraries the code reaches.
Hosted, and Foresight
Everything above runs free on your own machine. repowise.dev runs it for you, private repos included, and is where Foresight arrives first: models trained on years of real commits and fixes that rank the files a fix will touch, name what a change forgot, and explain every answer, with no LLM. What Foresight does →
Already indexed locally? repowise publish puts the repo on repowise.dev in one
command. It indexes your GitHub remote, so nothing is uploaded from your machine.
Public repos are free.
Know what breaks before you merge
All of this comes from the graph and git history, with no LLM, so the same diff always gets the same answer.
Change risk. Score any commit or branch 0 to 10 against your repository's own recent commits, with what it may break, the companion files it forgot, the tests it is missing and the tests to run.
repowise risk main..HEAD(reference)Tests without a coverage run. Which tests reach a file and which ones a diff exercises, from a coverage report when you have one and from the call graph when you do not: 97.5% correct on which tests to run.
repowise impacted-tests main..HEAD(reference)Bug magnets. Which files keep getting bug fixes, with a warning when your agent edits one. (reference)
Branch overlap. Other open branches editing the same files.
repowise overlap(reference)In CI. Patch coverage, doc drift, security and change risk as gates through the GitHub Action, a GitLab template or plain CLI, with SARIF output. No API key. Repowise in CI →
On every pull request. The free Repowise PR Bot posts one comment, edited in place on every push, with the callers outside the PR, the tests and companion files it is missing, and its change risk. A green PR gets no comment at all.
A real comment on a real PR: repowise-dev/repowise#1204 · its analysis page → · install the PR bot →
Every change, with its evidence
Each merged PR or commit becomes a Change Record: what it touched, who owns it, the tests that reach it, the contracts and decisions it crosses, and what happened after it landed. No forge API is called.
You ask | Repowise gives you |
What landed in this PR, and what backs it up? | The change's evidence checklist, each item |
Can CI hold changes to our payment code to a higher bar? | Rules in |
Something broke. Which change do we read first? | Recent changes ranked against an error, a stack trace or an issue. On 43 real regressions, the change a later revert or fix named was in the top 5 for 67% of them, against 28% for ranking by shared files. Suspects |
Who owns this, and are they still working on it? | Backstage-shaped components with owners from your catalog and CODEOWNERS, compared with the last 90 days of commits. Components |
Can we show an auditor or an acquirer where we stand? | Evidence packs: a repository assessment, a change report for a period and a component inventory, as JSON, Markdown, HTML or CSV. Reports |
Know exactly what to fix
Every file gets a 1 to 10 score from 53 deterministic detectors, read three ways: defect risk, maintainability and performance. Performance findings such as N+1 queries and I/O in loops are traced across functions and files.
The score is checked against real bugs. On 21 open-source projects, ranking by Repowise found 2.3x as many later-buggy files as CodeScene in the same review budget. It also checks itself on your repository after every index: "16 of the 20 lowest-health files had a bug fix in the last 6 months, 3.3x the 24% baseline."
Then it names the fix: Extract Method, Extract Class, Move Method, Split File or Break Cycle, with the exact symbols that move, ranked so a fix on a central file comes before the same fix on a leaf. Each plan has a copy-to-agent button.
repowise next # what to fix first, ranked by impact and effort
repowise health --refactoring-targets # concrete plans
repowise dead-code # what can go, by confidence tierCode health · Refactoring · Dead code · How it was validated →
Across your repositories
Real systems are more than one repository, and the expensive failures live between them. Change a backend contract and Repowise names the frontend calls that consume it, the services downstream, and the architecture rule the new dependency breaks, before it ships.
Workspace intelligence | What it answers |
Contract map | Which services provide and consume each HTTP, gRPC, event, socket and data contract? Links keep their exact or candidate confidence and the source evidence. |
Cross-repo blast radius | If this provider changes, which downstream services are in structural reach, and which may drift through historical co-change? |
Breaking-change guard | Was an endpoint removed or an OpenAPI, proto or signature shape changed incompatibly, and which consumer files are linked to that contract? |
Test impact | Which tests in the consumer repos should run for this provider change, measured from coverage or inferred from the call graph? |
Architecture as code | Does the live system graph violate declared dependency rules or contain cycles? |
Architecture health | How coupled is the estate? Propagation cost, the cyclic core, service roles and a deterministic 1-10 architecture score. |
Federated context | One dashboard and one MCP server answer across every repository while keeping repo-level evidence. |
Workspace guide and support matrix → · Keeping the index fresh →
Measured against the field
The same repositories, commits and questions for every tool, each given its full advertised surface, against answer keys we did not write.
Result | |
Work saved in an agent loop | 3.8 tool calls instead of 7.2, 3.0 files opened instead of 7.2, 31.6% less output. The best of five context tools. n=43, p<0.0001 |
Finding the right files | 0.876 of the files a fix needed, against 0.610 for the next tool, on a sealed set. n=42, p=0.00004 |
Command output | Up to 99.5% smaller ( |
Call graph the compiler agrees with | 93.8% to 99.8% precision on Go, TypeScript, Dart, C and Python, graded by each language's own toolchain. Against four other tools, none that finds as much of the graph gets more of it right, in 7 of 7 cells |
Health that predicts bugs | 2.3x CodeScene's recall of later-buggy files at the same review budget. 2,770 files, p=0.003 |
Scale | 15 million lines (dotnet/runtime) on one laptop. Building the call graph takes 75 MB, 10x less than the next tool |
Every result, in plain English, with the method → · The research it is built on →
No single product competes with all of this, so there is no single table. Rows marked measured are head-to-head numbers that link to docs/BENCHMARKS.md, where the sample sizes, the tests and the rows we lose live. Unmarked rows are capability presence, not measurements.
As an agent context layer
repowise | CodeGraph | Serena | DeepWiki | |
Self-hostable, open source | ✅ AGPL-3.0 | ✅ | ✅ | ❌ cloud only |
Private repo, no cloud | ✅ | ✅ | ✅ | ❌ OSS forks only |
MCP tools served | 10 default + 7 opt-in | 1 | 29 | 3 |
Finds the gold files (measured, n=42 sealed) | ✅ 0.876 | 0.610 | not in this run | not measured |
Output tokens vs a bare agent (measured, n=43) | ✅ -31.6% | -24.4% | -14.8% | not measured |
Memory to build the graph (measured, 5 tools, 35 repos) | ✅ 75 MB, lowest on 35 of 35 | 757 MB | not measured | n/a, cloud |
Time to build the graph (measured, same run) | 2.77s, fastest on 14 of 35 | 3.65s, fastest on 16 | not measured | n/a, cloud |
Time to build the full index, django (measured) | ⚠️ 366.8s, slowest here | ✅ 16.4s | not measured | n/a, cloud |
Call-edge precision, judged by a compiler (measured, 5 tools, 7 cells) | ✅ nothing that finds as much gets more of it right, 7 of 7 | lower precision in 7 | not measured | not measured |
Generated documentation | ✅ | ❌ | ❌ | ✅ |
Proactive agent hooks | ✅ Claude + Codex | ❌ | ❌ | ❌ |
Generated | ✅ | ❌ | ❌ | ❌ |
Command-output distillation | ✅ reversible | ❌ | ❌ | ❌ |
Architectural decision records | ✅ | ❌ | ❌ | ❌ |
Multi-repo workspace intelligence | ✅ contracts, co-change, federated MCP | ❌ | ❌ | ❌ |
The two cost rows answer different questions. Building the call graph, we are the lightest tool measured, about ten times lighter than the next, and roughly as fast as the fastest. Building the whole index, CodeGraph is 22x faster than we are, because by then we have also built the git-history layer, the wiki, the decisions and the health pass. If a call graph is all you need, that is the right trade and you should take it. With model-written prose on, it is 135x.
Competitors measured at CodeGraph 1.5.0, Graphify 0.9.31, Serena 1.6.2.dev0 and
code-review-graph 2.3.7 in August 2026. Repowise compiler-graded figures are from main
on 2026-10-03; the other Repowise rows are from 081a59fa, August 2026.
As a code health tool
repowise | CodeScene | |
Self-hostable, open source | ✅ AGPL-3.0 | ⚠️ on-prem Docker, proprietary |
Code health score (1-10) | ✅ 53 detectors, 25 scoring | ✅ 25-30 |
Brain Method / LCOM4 / god class | ✅ | ✅ |
Defects found at a 20% review budget (measured, 2,770 files) | ✅ 0.173 | 0.074 |
Effort-aware ranking, Popt (measured, p=0.003) | ✅ 0.607 | 0.462 |
Precision at that budget (measured) | 0.580 | ✅ 0.636, a shorter list |
Discrimination, ROC AUC (measured, paired) | 0.731 | 0.705, p=0.054, not significant |
Business impact (resolution time) | ❌ we could not replicate this on open data | ✅ Code Red study |
Git intelligence (hotspots, ownership, co-change) | ✅ | ✅ |
Pre-merge change-risk scoring | ✅ 0-10 + directives | ✅ |
Concrete cross-file refactoring plans | ✅ graph-aware + blast radius | ⚠️ within-function only |
Test-coverage intelligence | ✅ LCOV/Cobertura/Clover/JaCoCo/Go | ❌ |
Dead code detection | ✅ | ❌ |
Serves it to an AI agent over MCP | ✅ | ✅ |
CodeScene is the only other vendor in this category with a published empirical defect study, which is why it is the one we ran head to head. It flags about 27 files where we flag 132, so if you want a short list to act on, its threshold is the better fit.
Documentation generators
DeepWiki, Google Code Wiki and Swimm generate documentation from a repository, which overlaps one of our layers. We have not measured against them, so there is no table here.
The PR bot, against LLM review bots
Repowise PR Bot | CodeRabbit | Greptile | |
LLM calls per PR | ✅ zero | ❌ every review | ❌ every review |
Same diff, same review | ✅ deterministic | ❌ sampled output | ❌ sampled output |
Your code sent to a model provider | ✅ never | ❌ yes | ❌ yes |
Symbol-level blast radius | ✅ call graph | ❌ | ⚠️ prose, from context |
Co-change partners missing from the PR | ✅ git history | ❌ | ❌ |
Change risk vs the repo's own history | ✅ 0-10 + percentile | ❌ | ❌ |
Silent on a clean PR | ✅ by default | ⚠️ configurable | ⚠️ configurable |
An LLM reviewer is a different product: it can read intent, and it can be wrong in a new way on every run. This one does set arithmetic over a call graph and a git history, so pushing the same diff twice gives the same review twice. Full side-by-side comparisons: repowise.dev/compare →
Supported languages
28 languages parsed to an AST, 42 on a five-rung ladder, framework-aware where an ecosystem handler exists. Every language ships in the open-source distribution.
Rung | Languages | What you get |
Full (15) | Python · TypeScript · JavaScript · Svelte · Vue · Astro · Java · Kotlin · Go · Rust · C++ · C# · Scala · Ruby · Apex (Salesforce) | The whole pipeline: AST symbols, import resolution, a resolved call graph, heritage, docstrings, framework edges and code-health markers |
Good (11) | C · Swift · PHP · Dart · Object Pascal · COBOL · GDScript · VB.NET · Elixir · F# · Objective-C | All of the above except the full health suite, within the language-specific ceilings in the full matrix |
Partial (2) | Luau / Roblox · Razor / Blazor | Luau: AST symbols and |
Lightweight (6) | Clojure · Haskell · Lean 4 · Erlang · HTML · QML | A real file-to-file import graph, and no symbol-level claims |
Structural (8) | R · Zig · Julia · Elm · OCaml · Crystal · Nim · D | Git history: blame, hotspots, co-change, ownership, bug history |
SQL and dbt projects get ref() / source() lineage, shell scripts get function-level
symbols, HTML pages contribute their <script src> and <link href> dependencies, and
OpenAPI, Protobuf, GraphQL, Dockerfile, Terraform and similar formats get dedicated
handlers. Anything else is still tracked through git history.
Every call edge is stamped with how it was resolved and how much to trust it, from
same_file at 0.95 down to a repo-wide name match at 0.50, labelled as the guess it
is. Accuracy per language, graded against each language's own compiler:
docs/BENCHMARKS.md#accuracy-by-language.
Full matrix: docs/layers/LANGUAGE_SUPPORT.md → · adding a language takes five small steps and no changes to the parser core: docs/architecture/language-support.md → · languages moving up the ladder: roadmap →
Supported agents and editors
Nine agents wired end to end · two at the Full tier · every other MCP host one paste away.
Full is every surface Repowise has: MCP tools, skills, slash commands, a managed
instructions file, hook-level interception of tool calls, and transcript mining after
the session. Good is MCP tools plus an instructions file or skills, without transcript mining
(hooks only where the matrix says). Basic is the MCP config alone. Anything else that speaks MCP is one snippet away:
repowise agents print-config claude-code prints a server entry for Cline,
Zed, Gemini CLI or any host that reads mcpServers.
Integration matrix →
In VS Code, the Repowise extension shows what your change breaks before you push (riskiest files, what is downstream, forgotten companion files, missing tests, suggested reviewers), health in the gutter and status bar, callers and ownership on hover, and refactoring plans as CodeLens. One install also registers the MCP server, so the same index serves you and your agent. Install from the Marketplace or Open VSX and run Repowise: Set Up This Repository. VS Code guide →
In Claude Code, Lens ships inside the Repowise plugin and shows you what the index
knows while Claude works: the file it is on and how many files depend on it, a change
review after a turn that edits files (code health, tests to run, other branches on the
same files), and /lens, a pane with Flow (what to check before accepting each turn),
a map of the repo lit by what Claude searched, opened and edited and what its edits
reach, Ask, and a session recap. Nothing reaches Claude unless you press a button, and
Lens makes no model calls of its own. It needs Claude Code 2.1.287 or later, and the
map needs repowise serve --no-ui running. Lens guide →
The ten MCP tools
Every response carries a _meta envelope with the indexed commit, the index age and a
stale warning when the index has fallen behind your checkout, so your agent always
knows how much to trust what it just read.
Tool | What only this tool answers |
| Architecture summary, module map, entry points, git health. The first call on an unfamiliar codebase. |
| Hybrid retrieval (full-text plus vector), graph expansion and one cited answer with a single calibrated |
| Triage card for files, modules or symbols: summary, signatures, hotspot flag, governing decisions, symbol ids. |
| One indexed symbol's source with exact line bounds. |
| Hybrid search over code and docs, by symbol, path or concept. |
| Hotspots, dependents, co-change partners, ownership, test gaps, bug history. Pass |
| What a commit, range or uncommitted change made worse across defect risk, maintainability and performance, the tests that touch it, and how the diff ranks against recent commits. |
| Architectural decisions with their verbatim evidence. Falls back to git archaeology when no decision exists. |
| Unreachable code by confidence tier, with cross-repo consumers in workspace mode. |
| Health scores and findings across all three lenses, Fix first, coverage, trends, doc drift and refactoring plans. |
Ten is a deliberate ceiling: a small, task-shaped surface is easier for an agent to choose from than a large one. Seven more tools (dependency paths, execution flows, refactoring code generation, finding triage, and three workspace architecture tools) are opt-in. Parameters, examples and when to use which: docs/agent/MCP_TOOLS.md →
For teams and enterprises
AI makes producing a change cheaper. It does not make understanding its consequences cheaper. Repowise gives developers, agents, reviewers and platform teams the same evidence about what exists, what depends on it, what is risky and what will break, from one index, in place of a separate health tool, dead-code tool, code-search layer for agents, docs generator and test-impact service.
Deployment | Where your code is read | What leaves your network |
Self-hosted, open source ( | your machine | anonymous CLI telemetry you can switch off, and your own LLM provider only if you turn prose on |
Self-hosted, commercial | your VPC or an air-gapped network | nothing at all in air-gapped mode |
Hosted (repowise.dev) | our managed indexer | your code goes to the platform |
Graph, git, health, change risk, tests, dead code and PR review make zero LLM calls, and raw source is never stored. Prose is optional and runs on your own provider or fully offline through Ollama.
Evidence you can hand to someone else.
Who is asking | What Repowise produces |
An auditor reviewing change management (SOC 2 CC8.1, ISO 27001 A.8.32, PCI DSS 6.5) | Every change in the period with its review record, reaching tests, policy verdict and later outcome: |
A software inventory review (PCI DSS 6.3.2) | Components with owner teams, languages, entry points and the packages they import: |
A security team asking which of your code uses a package | Import sites with component, owner team and reachability, plus a CycloneDX 1.6 usage export: |
An acquirer, a board or a migration committee | Component posture, hotspots, bus factor, dead code, test reach and contracts: |
An incident review | The recent changes to read first, with the signals behind each: |
Any git host works, because indexing reads a local checkout: GitHub (PR bot, Action, webhooks), GitLab (CI template, webhooks), Bitbucket and others through the CLI, and plain directories or Perforce and SVN exports for every layer except history.
Commercial licences come per seat, per indexed repository or enterprise-wide, with a named contact, an SLA and a quarterly architecture review. Using Repowise inside your company under the AGPL creates no obligation to publish your own code.
Commercial detail and capability status · Security review pack · Evaluating it for your company · hello@repowise.dev
Privacy
Deterministic or offline mode: with
--no-prose, code-derived content stays on your infrastructure. The CLI reports anonymous, opt-out usage telemetry (command names and coarse environment only); turn it off withrepowise telemetry disable,DO_NOT_TRACK=1, or by running fully offline. What's collected →Optional LLM features: generated prose, decision extraction and refactoring code generation send code-derived prompts directly to the provider configured with your own key. Repowise does not proxy those calls.
What's stored: the graph, embeddings, generated wiki pages and git metadata. Raw source is processed transiently and never persisted.
Fully offline: Ollama plus a local embedding model means zero external calls.
Doing a security review? docs/business/SECURITY_COMPLIANCE.md →
CLI
repowise init [PATH] # index a codebase (plan screen; --yes --no-prose needs no key)
repowise generate [PATH] # answers and docs with a model, on an existing index (no re-index)
repowise serve [PATH] # MCP server + local dashboard
repowise update [PATH] # incremental update (--workspace for every repo)
repowise watch # re-index on file change
repowise search "<q>" # hybrid search (fulltext / semantic / symbol / path)
repowise ask "<q>" # a synthesized answer with citations
repowise context <files> # triage card: layer, hotspot, fix history, freshness
repowise symbol <id> # one symbol's body, with verified line bounds
repowise why <q|path> # decisions, rationale, git archaeology
repowise next # what to fix first
repowise health # code-health KPIs and lowest-scoring files
repowise risk main..HEAD # score a branch or PR range
repowise overlap # other branches editing the same files
repowise impacted-tests # only the tests a diff exercises
repowise changes show ID # one merged change and its evidence checklist
repowise check # evaluate .repowise/policies.yaml against a change
repowise suspects # recent changes ranked against a symptom
repowise components # declared and derived components, owners vs activity
repowise packages where # where a third-party package is imported
repowise report # assessment, change evidence and inventory packs
repowise dead-code # unreachable code by confidence tier
repowise doc-drift # documentation the code no longer supports
repowise security # secrets and risky patterns, tree or full history
repowise decision list # architectural decisions
repowise export --format structurizr # the architecture as Structurizr DSL
repowise distill pytest # compact, errors-first, reversible command output
repowise saved # tokens and dollars saved by distillation
repowise workspace add # multi-repo workspace management
repowise doctor # check setup, API keys, index drift
repowise publish # put this repo on repowise.dev (indexed from GitHub; public repos free)
repowise uninstall # remove what repowise wrote, and say what it leftEvery command and flag: docs/reference/CLI_REFERENCE.md · config: docs/reference/CONFIG.md · something not working: docs/start/TROUBLESHOOTING.md · all docs: docs/README.md
Contributing
git clone https://github.com/repowise-dev/repowise
cd repowise
uv sync --all-packages
uv run repowise --version
uv run pytest tests/unit/New here? You do not have to read 3,000 files to start. We keep a public index of this repository built by Repowise itself, re-indexed on every push: explore repowise with repowise → (architecture, hotspots, ownership, decisions, and a ranked refactoring backlog you are welcome to pick from).
Full guide, including how to add languages and LLM providers: CONTRIBUTING.md · architecture: docs/architecture/
License
AGPL-3.0. Free for individuals, teams and companies using Repowise internally.
For commercial licensing (the enterprise security and compliance layer, SSO/SCIM, RBAC, workflow integrations, priority support and SLA, or embedding Repowise in a product without AGPL obligations), see docs/business/COMMERCIAL.md or contact hello@repowise.dev.
Built for engineers who got tired of watching their AI agent cat the same file for the fourth time.
Available Tools
5 toolsget_answerARead-onlyInspect
Answer a how, where, or why question in one evidence-grounded call.
High confidence may be used directly; medium and low lead with the
evidence to verify. ``candidate_files`` ranks files worth opening that
the citations do not name.
Args:
question: One question about one subject. Split "where is X and
when does Y run" into two calls; do not paste a whole
multi-part task.
scope: repository-relative path prefix.
repo: usually omitted.
include: ``["evidence"]`` expanded evidence; ``["diagnostics"]``
adds ``retrieval_quality``.
Inputs: question (required), repo, scope, include. Returns: result
get_contextARead-onlyInspect
Triage card for files / modules / symbols — relationships, not source bytes.
Returns summary, signatures with lines, hotspot, fix_history and
decision titles; either risk cue means get_risk. Batch targets in one
call. Whole file, body-elided and line-verified: include=["skeleton"]
in ONE call, or Read it; do not call get_symbol per signature. Rename:
include=["references"] is the whole edit set when ``complete``. Which
files in a directory lack tests: include=["files"] on it.
Args:
targets: paths, "path::Symbol", "component:<name>",
"change:<id>" (pr-123 or a sha), "package:<name>".
include: callers | callees | references (+ adds comments) | skeleton | skeleton+
| files | symbols (all, not top 15) | decisions | full_doc
| ownership | last_change | metrics | community | health
| doc_drift | changes | provenance | suspects | facts.
compact: False adds structure, imports, docstrings.
symptom: issue text or stack trace to rank ``suspects`` on.
repo: usually omitted.
Inputs: targets (required), repo, compact, include, +1 more. Returns: result
get_symbolARead-onlyInspect
Follow-up read of one symbol whose id another response already gave you.
**Not an entry point.** ``get_answer`` ships ``symbol_bodies``; for a
whole file ``get_context(include=["skeleton"])`` or a Read is one call.
Never walk a file symbol by symbol. Returns verified, line-numbered
source; a truncated result carries the ``continuation`` to pass back.
Args:
symbol_id: "path/to/file.py::Name", "path/to/file.py:140-180" for a
live range, or an omission ref. A bare file path is not an id.
depth: 1 (default) is this symbol alone; 2-3 also returns the bodies
it calls, transitively, in ``callee_bodies``.
context_lines: extra lines before/after (0-50).
query: regex/substring filter on an omission ref's lines.
id, reference: aliases for ``symbol_id``.
repo: usually omitted.
Inputs: id, repo, depth, query, +3 more. Returns: result
get_whyARead-onlyInspect
Why this code is shaped this way — decision records + evidence commits.
Call before a refactor or pattern divergence. Query: a question, a file
path, a question with ``targets``, or nothing (decision dashboard).
A path gets the rules in force first, then labelled background records.
``authority: accepted`` means somebody signed the decision;
``answer_basis`` names the strongest evidence lane, and only
``decision`` is a ruling.
Args:
query: question, file/module path, or omit for the dashboard.
targets: file paths to anchor the search.
id: decision or ``ev_...`` evidence id from any tool.
reference: structured evidence reference.
repo: usually omitted.
include: ``["history"]``: a path's origin story, commits, episodes.
Inputs: id, repo, query, include, +2 more. Returns: result
search_codebaseARead-onlyInspect
Find code by concept, symbol, or path — hybrid codebase search.
For a how/where/why question call get_answer. Use this to enumerate
matches, resolve an identifier to a symbol_id, or find files for
get_context. Identifier, literal, `a|b|c` or regex queries return
`lines` (path, line, kind, text; definitions first, one group per
alternative) read from live files; when `complete` is true they are
every match, so no grep is needed. `candidates` lists
distinct files to Read, best first.
Args:
query: identifier, `a|b` alternation, regex, path, or natural language.
limit: max results (default 5).
mode: auto (by query shape) | symbol | path | concept | hybrid.
kind: implementation | test | config | doc.
symbol_kind: function | class | method | ...
page_type: file_page | module_page.
repo: alias, or "all" for workspace-wide.
pattern: alias for query.
Inputs: kind, mode, repo, limit, +4 more. Returns: result
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 5 tools
The core intents (search, triage context, symbol body, question answering, decision history) are broadly distinct, but the descriptions spend large amounts of text telling the agent which sibling *not* to use ("Not an entry point", "For a how/where/why question call get_answer"). get_answer and get_why both claim the "why" space, and get_context vs get_symbol both return source-ish content, so boundaries are guidance-dependent rather than intrinsic.
Four of five tools follow a clean get_* verb_noun pattern and the fifth (search_codebase) is a standard verb_noun variant for its query-based role. Minor deviation only, still highly readable and predictable.
Five tightly focused tools, each mapped to a distinct retrieval mode (question, search, triage, symbol read, decision recall). Nothing feels padded or redundant in raw count for a code-intelligence server.
Retrieval lifecycle is well covered: discovery (search_codebase), orientation (get_context), deep read (get_symbol), rationale (get_why), and synthesized answers (get_answer), with cross-tool chaining via symbol_id/omission refs. Gaps are limited to no repository/workspace enumeration entry point and read-only scope (no edit or index-refresh operation), which agents can mostly work around.
Maintenance
Related MCP Connectors
Code intelligence platform for AI agents. 20 tools for architecture, security & impact analysis.
Codebase knowledge graph — architectural awareness for AI coding agents via MCP
Codebase intelligence for agents: 152 structured artifacts across 21 programs, one call.
Team memory for AI coding agents and code reviews, reviewed by humans and served over MCP.
Related MCP Servers
- AlicenseNot gradedqualityDmaintenancePersistent codebase knowledge layer for AI agents. Pre-digests codebases into structured knowledge (symbols, dependency graphs, co-change patterns, architectural decisions) and serves via MCP. 28 languages, 14 tools, ~85% token reduction.22 npm10MIT
- AlicenseNot gradedqualityCmaintenanceEnables AI agents to search code by meaning, explore codebase structure, store and query knowledge with temporal facts, and read source code through a set of MCP tools.7MIT
- AlicenseNot gradedqualityCmaintenanceProvides code intelligence for AI coding agents by indexing repositories into a hybrid knowledge graph, enabling agents to query dependencies, impact, and context through 28 MCP tools.4Apache 2.0
- AlicenseNot gradedqualityBmaintenanceProvides AI agents with a function-level dependency graph of the codebase through 30 MCP tools, enabling structural queries about code dependencies, callers, and impact analysis.2,524 npm101Apache 2.0