Comments
>log in to commentUnlike traditional line‑by‑line PR views, /pr‑viz turns a GitHub pull request into a two‑zone map that highlights blockers and AI‑cleared items. I built it to cut review time and to work in self‑hosted, air‑gapped environments using a simple skill you call from your IDE. Let me know what works and what could be improved 😊
Free IDE skill that turns a GitHub pull request into an interactive two-zone review map: needs-review vs AI-accepted. Run /pr-viz from Cursor, Claude Code, VS Code, or Windsurf — the map opens at pr-viz.com or on your own host.
Built for self-hosted and air-gapped setups: Cosign-signed Docker viewer, no required SaaS. Review what matters, filter the noise, optionally explain/fix in the IDE and post comments as yourself from the browser.
Sample map: pr-viz.com/d/sample · Skill: github.com/krzysztofgibert/pr-viz-skill
- Self‑hosted, air‑gapped PR review maps
- Runs on your network with no internet required
- Integrates with IDEs via skill calls
- Reduces review time by focusing on blockers
- No GitHub token stored on the server
Free IDE skill that creates an interactive two-zone review map for GitHub pull requests.
- for
- Developers and teams needing self-hosted PR review visualisation.
- pricing
- free
- license
- MIT
Key features
- Two-zone review map — Shows needs-review vs AI-accepted sections of a PR.
- Self-hosted & air-gapped — Runs on your own Docker host with no internet required.
- IDE integration — Triggered from Cursor, Claude Code, VS Code, Windsurf via /pr-viz command.
- Comment posting as you — Posts AI-generated comments to GitHub using your credentials.
- Human-readable AI notes — Provides short explanations and fixes for highlighted issues.
- PR walkthrough — Animated spoken summary of verdict, issues, and failing checks.
Use cases
- Start review with blockers and skip AI-cleared noise.
- Share a consistent risk diagram with the whole team via a URL.
- Generate inline GitHub comments from IDE without extra setup.
- Run PR analysis in a secure, private network without SaaS.
pr-viz vs alternatives
 pr-viz |  Qodo |  Augment Code |  Tabby ML |  Continue.dev | |
|---|---|---|---|---|---|
| Best for | Self-hosted PR review maps | AI code review & test generation | Context-aware AI coding | Self-hosted AI coding assistant | IDE AI assistant & CI checks |
| Pricing | Free | Subscription | Subscription | Free | Free |
| DevHunt upvotes | 1 | 0 | 0 | 0 | 0 |
| Launched | Nov 2026 | — | — | — | — |
- pr-viz vs Qodo: Focuses on AI code review and test generation, not visual PR maps.
- pr-viz vs Augment Code: Provides deep codebase context for AI assistance, not a self-hosted review map.
- pr-viz vs Tabby ML: Self-hosted AI coding assistant, but not specialized for PR visualisation.
- pr-viz vs Continue.dev: Open-source IDE assistant with checks, not a PR map visualiser.
pr-viz FAQ
Do I need an internet connection to use /pr-viz?+
No, the viewer and API run on your host and can operate completely offline.
Which IDEs can trigger the skill?+
Any IDE that can call a skill, including VS Code, Cursor, Claude Code, and Windsurf.
Can I use my own LLM for the analysis?+
Yes, /pr-viz leverages the model you already pay for; it does not host an LLM.
How are comments posted to GitHub?+
The tool copies the comment and opens the exact line on GitHub, posting as your user.
Is the Docker image signed?+
The viewer is Cosign-signed for supply-chain security.
What does the ‘air-gap’ capability mean?+
After installation the service runs without outbound internet, keeping maps inside your network.
Summarized by DevHunt from pr-viz.com · Oct 27, 2026. Details may change; check the official site.
