PR Explainer doesn't just throw your diff at an LLM and hope for the best. It builds a context graph that resolves dependencies, traces call sites, and classifies changes by severity, so every tour and review references real code at verified line numbers. The narrated walkthroughs explain what changed; the review findings flag what's wrong. Both read actual code, not a few lines of a patch.
How it works
Any PR, any size. Run "Explain PR" from the command palette.
A voice walks you through each change while the relevant lines light up. Pause, skip, or rewind.
See severity-graded findings inline while you listen. Actually understand what you're approving. Export as video for the rest of the team.
Features
Playback
The narration walks through each change while the relevant lines light up. No context switching, no re-reading diffs.
Code Review
Severity-graded findings (critical bugs, warnings, good practices) shown inline with gutter markers in the editor. Run reviews standalone or combined with narrated tours.
Export
Export a tour as an MP4 with captions or a small, selectable-text PDF. Review findings travel with both, so reviewers see the flags inline.
Share
Export a tour as a portable .prtour file and share it with anyone. It opens against their checkout — live where the code matches, re-anchored where it moved, and the author's snapshot as a fallback. No repo access required.
Detail Levels
Quick 30-second overview or a thorough file-by-file breakdown. You pick how deep to go.
Smart Caching
Tours are cached by content hash. Revisit a PR and it plays back instantly. No regeneration, no waiting.
Multiple Providers
Claude, GPT, or Gemini: use whichever LLM you prefer. Managed mode picks the best available; local mode uses your own API keys.
Daemon mode
Leave the background daemon running and your work turns into a narrated chronicle: a readable history of why the code changed, not just what. It flags the notable moments and pins them. There's no PR to open and no prompt to write; it builds up quietly while you work, so onboarding, standups, and “what changed last week?” get a lot easier to answer.
A run of commits lands a token-exchange flow so people can sign in with an account they already have, the groundwork for single sign-on across the suite.
142 commits from 3 contributors converge: the public API surface is now frozen and versioned, and the rollout path is documented end to end.
A guard closes a crash where a second logout hit an already-cleared session. Now a quiet no-op instead of a panic.
Per-route throttling is pulled into middleware, so every new endpoint is protected by default instead of opt-in.
Every change, narrated in plain language. Scroll the whole history of the repo.
Releases, big refactors, and turning points detected and pinned automatically.
Repo health and momentum at a glance. Runs locally, so your code never leaves your machine.
Pricing
Also runs as a CLI (prexplainer review). Installed and
kept up to date by the Indieprise suite
(Docker or ids install prexplainer). One indieprise account
and API key works across all tools.
Want unlimited tours? Join the Pro waitlist
See it in action
An actual tour of a real commit · Open in new tab ↗
A real walkandtalk tour of a real commit: the Windows port of its own daemon
Inside your editor — the main way to use prexplainer.
Or install from the VS Code Marketplace / OpenVSX (Cursor, Windsurf). Sign in from the extension, or paste an API key from your settings.
One-off reviews and tours from the terminal, plus CI.
Run only prexplainer's daemon in a container — isolated, no local install.
The whole portfolio in one container — prexplainer alongside migrataur & devsbill.
Always-on: your repo narrates itself as a Code Story, with git-event reviews on your schedule.
One indieprise account works across prexplainer, migrataur, devsbill, and the suite dashboard. Generate keys at prexplainer.com/settings. A key created here also authenticates every other indieprise tool.
Take back control of code review. Understand every change before you hit approve.