Intraview – AI Code Tours Inside VS Code

TL;DR

TL;DR: Intraview is a VS Code extension that generates interactive, step-by-step code walkthroughs from your actual codebase using any AI agent — local-first, git-stored, with self-healing when code drifts.

What Is Intraview?

Intraview is a VS Code extension that bridges the gap between AI code generation and human understanding. When you ask your AI agent to explain or modify code, Intraview converts that into a structured, navigable tour that your team can follow, refine, and store in git.

The core premise: AI agents generate code constantly, but the explanations stay trapped in chat threads. Intraview externalizes those explanations as first-class artifacts in your repo.

From the product description on intraview.ai:

“AI creates code tours. You refine them visually. Your team onboards faster. Local-first, git-stored, works with any AI agent.”

Key characteristics:

• VS Code extension (Visual Studio Marketplace)
• Tours stored as JSON files in your repository
• No cloud account required — code never leaves your machine
• Works with any AI agent that can read and write files
• Detects code drift and prompts tour updates automatically

Source and Accuracy Notes

⚠️ This section is MANDATORY. All links must be verified from actual source, not guessed.

• Project page: intraview.ai — verified product description, OG image
• VS Code Marketplace: marketplace.visualstudio.com/items?itemName=intraview.intraview — extension install page
• HN launch thread: news.ycombinator.com/item?id=45837067 — Show HN with 48 points
• License: not yet verified from source — check repo before making claims

How It Works

Intraview’s workflow has three stages:

Step 1: Ask

Tell your AI agent what you want to understand — “Walk me through the auth system” or “Explain the payment flow.” The agent reads the relevant files, identifies key sections, and builds a tour structure.

Step 2: Refine

The generated tour opens directly in VS Code. Click any title, description, or step to edit inline. No JSON editing, no config files. What you see is what your team gets.

Step 3: Share

Tours are saved as JSON files in your repository. A git push distributes them to your team. When code changes, Intraview detects the drift and prompts you to update the tour — no silent failures.

Practical Evaluation Checklist

Onboarding

• Works with any codebase — no special framework required
• AI generates the initial tour; human refines for accuracy
• New hires get navigable walkthroughs instead of “just read the code”

Code Review

• Agent builds a guided tour of a PR diff
• Reviewer leaves inline feedback at each step
• Tour becomes part of the review record in git

Agent Collaboration

• Agent demonstrates its changes through a structured tour
• Human provides feedback; agent refines based on annotations
• Export annotated feedback as context for the agent

Self-Healing Tours

• Code moves; tours break silently — Intraview detects drift when underlying code changes
• Prompts update rather than leaving stale explanations
• No cloud dependency to track drift

Privacy

• Local-first: no account, no cloud sync, code never leaves your machine
• Tours are JSON in your repo — same access controls as your code

Setup

Intraview is a standard VS Code extension. Search for “Intraview” in the Extensions panel or install directly from the Visual Studio Marketplace.

After installation, open a project and invoke Intraview from the command palette. The extension will prompt for an AI provider — configure your preferred API key or use a local model endpoint.

FAQ

Q: Does it work with any AI agent? A: Yes — Intraview reads and writes files, so it works with any agent that has file access (Claude, GPT, Cursor, Copilot, etc.). The extension handles the tour format; the agent handles the reasoning.

Q: Where are tours stored? A: As JSON files in a .intraview/ directory (or your chosen path) within your repo. They travel with your code via git.

Q: Does it require an internet connection? A: The extension itself needs to reach your configured AI provider. If you use a local model (Ollama, etc.), Intraview works fully offline.

Q: How does drift detection work? A: Intraview tracks which files and line ranges a tour references. On open, it checks whether those locations still match. If code moved or was deleted, it flags the affected steps and prompts you to update.

Conclusion

Intraview solves a real problem in AI-augmented development: explanations disappear into chat threads. By converting AI-generated code understanding into git-stored, version-controlled tours, it makes that knowledge permanent and team-accessible.

The local-first, no-account model is the right default for a dev tool — privacy without tradeoffs. If your team uses AI agents for code review, onboarding, or collaborative debugging, Intraview is worth a look.

Install it from the VS Code Marketplace and run intraview from the command palette to start your first tour.