Give any AI instant understanding of your codebase — no explaining, no copy-pasting, no wasted tokens.
Primer indexes your project into a live dependency graph, tracks your active session, and generates ready-to-use context for Claude, Cursor, ChatGPT, Gemini, or any AI tool. Open the sidebar, click "Generate briefing," paste. Done.
How it works
Primer runs locally. No API key. No cloud. Your code never leaves your machine.
- Install the extension and open your project folder
- Initialize — Primer scans your codebase and builds a symbol dependency graph
- Watch — Start a session to track which files and clusters you're working in
- Paste — Generate a briefing and paste it into any AI chat as your first message
The AI immediately understands your architecture, what you were working on last session, and what constraints apply — without you writing a single word of explanation.
Features
Live Sidebar Dashboard
Always-visible panel showing your session status, architecture clusters, and active constraint rules. Updates in real-time as you work.
Architecture Brief (Tier 1)
Primer ranks every symbol in your codebase by architectural importance (in-degree, churn rate, cluster centrality) and surfaces the top 20 as a ~500 token brief. The exact right amount of context — not too much, not too little.
Session Tracking
primer watch logs which files and clusters you touch as you code. When you stop, a session echo is written. Next AI session: paste the echo → the AI knows exactly where you left off.
Generate AI Briefing
One click. Gets you:
- Tier 1 architecture brief (top symbols + scores)
- Session echo (files touched, last cluster)
- Current file preview
- Active constraint rules
Copy it. Paste it. Your AI is fully briefed.
What Depends on This?
Right-click any symbol → see every function that calls it and every function it calls. Know the impact before you change anything.
Veto Warnings
When you open a file in a cluster with active constraint rules, Primer shows a warning automatically. No more accidentally reverting critical changes.
Impact Analysis
Place your cursor on any symbol → Primer: What Depends on This? → get a full report of callers and dependencies from the live graph.
Commands
| Command | What it does |
|---|---|
Primer: Initialize |
Scan codebase, build dependency graph |
Primer: Start Watch |
Track live session (files + clusters) |
Primer: Generate AI Briefing |
One-click context for any AI |
Primer: What Depends on This? |
Impact analysis from cursor position |
Primer: Export Context |
Write full context to primer-context.md |
Primer: Copy Tier 1 |
Copy architecture brief to clipboard |
All commands available from the sidebar buttons — no command palette required.
Works with
- Claude (claude.ai, Claude Code, Cursor)
- ChatGPT / GPT-4
- Gemini (Google AI Studio, Antigravity)
- Cursor (via sidebar briefing)
- Any AI that accepts text input
Primer does not call any AI API. It prepares context for the AI of your choice.
Requirements
- Node.js 18+
- A project folder open in VS Code
Run npx primer-ctx init once in your project terminal to initialize. The sidebar populates automatically.
Changelog
0.1.2
- Embedded dashboard view in sidebar (no floating tab needed)
- All actions accessible from sidebar buttons
- Right-click context menu for impact analysis and briefing generation
- VS Code 2.x compatibility
0.1.1
- New graph-node icon
- Veto file warnings
- Impact analysis command
- AI Briefing generator
0.1.0
- Initial release: session tracking, architecture map, tier 1 brief, export