Written as you push
Every push to an open pull request, Flash drafts a description for any public symbol that's still missing one, batched per changed file, so docs never fall behind the code.
Evidence-driven code intelligence
AI code intelligence that has to show its work.
Aletheore reviews, audits, and monitors repositories with every finding traced back to source evidence: file, line, symbol, owner, commit, dependency, and risk.
Source-availableLocal-first CLIHosted GitHub App
Every alert, review, audit, and query resolves back to:filelinesymbolownercommitdependencyrisk
How resolution works
Not "it broke." A chain you can walk.
Every alert Aletheore raises is a resolved path, not a guess. This one is real: it was read from a scan of Aletheore's own repository.
- endpointGET /admin/{org}/{repo}
A route the scanner found in github-app/app_server/admin.py.
- handleradmin.py:540
The scanner resolved the route to github-app/app_server/admin.py at line 540.
- symboladmin_page
admin_page spans lines 541 to 601 of github-app/app_server/admin.py.
- ownerArihant Kaul
49 commits touch this file, 100% of them by this author.
- last touchedfca0676
2026-09-25: restructure: Overview page and dashboard shell to match the approved mockup
- verdictHotspot: 49 commits
49 commits and 3 dependent modules put this file on the hotspot list.
What it covers
One evidence layer. Four useful surfaces.
Run locally with the CLI, install as a GitHub App, or give coding agents a smaller evidence context through MCP.
Code Review
Automatic PR comments and Flash reviews on every push, plus managed audits and branch-protection checks that cite the changed file and line.
- Review only what changed
- Catch secrets before merge
- Keep findings source-grounded
AIRview, live
This is a real dependency graph, not a screenshot.
AIRview builds this from your own repo's real imports: drag a node, hover to trace what depends on what. This demo is graphed from Aletheore's own module names.
Hover a module to see what it imports.
Why it is different
Most AI tools summarize. Aletheore resolves.
Aletheore starts with deterministic repository evidence, then lets AI reason only over what the scanner can prove. The result is friendlier than raw static analysis and more trustworthy than a generic AI summary.
Docs
Nobody wants to write the docs. So nobody has to.
Every team says "we'll document it later," and later never comes: the code always outruns the wiki. Aletheore AIR writes the docs instead: grounded, per-symbol descriptions generated straight from your real source, kept current automatically, with zero engineer hours spent typing a docstring nobody wanted to write. Export the whole set as a single Markdown file whenever you need it outside the dashboard.
How Docs stays honest
Generated from your code. Never invented.
Same evidence-grounded discipline as every other Aletheore surface: AI only describes what the scanner can prove exists.
Fully documented on day one
The moment you connect a repo, AIR runs a full first pass across everything already public, so you're not starting from a blank docs page.
Marked, never mistaken for yours
Every AI-touched description is grounded in the real source snippet and carries an explicit AI-generated or AI-polished marker, never silently presented as something a developer wrote.
Self-healing by design
Progress is saved symbol by symbol, so one failing file never costs the ones that already succeeded. A 48-hour sweep automatically finishes any repo with real commit activity since its last pass.
How teams use it
Start local. Then let the GitHub App keep watch.
$ aletheore scan
Build the evidence
Tree-sitter and git history produce air.json and air.toon. No LLM in the scan path.
$ aletheore query
Ask grounded questions
Query architecture, owners, dependencies, endpoints, secrets, hotspots, and semantic search from one evidence file.
on: pull_request
Review every change
The GitHub App comments on PRs, gates new secrets, and can run managed audits or Flash reviews on paid plans.
GET /health
Monitor what shipped
Live endpoint checks alert on outages and latency regressions with source handler context.
Install the GitHub AppSee PricingInstallable today on any GitHub account or organization you administer.
Proof
Scanned for real. Not a mockup.
Three real public repositories, actually scanned by aletheore scan at pinned commits. The numbers below are read directly from that output. Read the full writeup, findings and all →
Django
@ 3d34265Python, full git history
- modules
- 3,038
- dependency edges
- 9,479
- clusters
- 896
- commits scanned
- 34,793
- license findings
- 0
- vulnerability findings
- 1
Express
@ ae6dd37JavaScript, real npm dependencies
- modules
- 141
- dependency edges
- 0
- clusters
- 141
- commits scanned
- 6,157
- license findings
- 0
- vulnerability findings
- 0
Kubernetes
@ bd1a1b8Go, the stress test
- modules
- 17,499
- dependency edges
- 59,174
- clusters
- 12,641
- license findings
- 40
- vulnerability findings
- 9
- seconds to scan
- 368
Real stress test
The Linux kernel.
Not a curated sample: the actual torvalds/linux repository, cloned in full and scanned end to end: 94,849 files, 8.2GB, 1.46 million commits, the latest 50,000 of them analyzed. A re-scan takes 2m 29s.
- Files parsed
- 64,554
- C files
- 36,846
- C++ / header files
- 26,867
- Dependency edges
- 71,923
- Architecture clusters
- 27,266
- Commits analyzed
- 50,000
- License issues found
- 0
- Known vulnerabilities
- 0
- Scan time
- 5m 6s
The same data, TOON-encoded.
Measured against Aletheore's own air.json / air.toon pair from a real self-scan (2026-09-25, o200k_base tokenizer), the exact data the MCP server and audit coding-agent adapter read.
47.2%
fewer tokens
53.8%
smaller on disk
How it measures up against a real competitor, wins and losses both: See the full benchmarks →
Humanist tools for a technical world.
The deterministic local scan does not upload your code. The CLI sends a small crash report only if it crashes, and one command turns that off. Bring your own API key for AI-assisted audit reports, or skip them entirely: the evidence pipeline stands on its own.
- →Source-available and local-first.
- →Query, diff, dashboard, MCP, and reports all read the same evidence.
- →Hosted paid-plan features use derived evidence, not raw source storage.
$ aletheore scan .
Scanning /Users/you/your-repo...
Detecting languages, frameworks, and build tools
Building module dependency graph
Checking dependency vulnerabilities
Evidence written to .aletheore/air.json
Snapshot saved to .aletheore/history/