Quick Start
Get Code Insights running in 30 seconds. One command syncs your AI coding sessions and opens the local dashboard. No accounts or API keys required.
Try it instantly
npx @code-insights/cliNo installation needed — try it right now.
Or install globally
npm install -g @code-insights/cliThen run a single command to sync your sessions and open the dashboard:
code-insightsThat's it. Code Insights auto-detects your AI coding tools (Claude Code, Cursor, Codex CLI, Copilot CLI, VS Code Copilot Chat), syncs all sessions into a local SQLite database, and opens the dashboard at localhost:7890.
The first run processes your entire session history and may take a minute. Subsequent runs are incremental.
Set up auto-sync (optional)
Install a hook that runs a quiet sync whenever a Claude Code session ends — so your dashboard is always up to date without thinking about it:
code-insights install-hookGet AI insights
Claude Code users get zero-config analysis — the hook you just installed analyzes every session automatically using your Claude subscription. No API key needed. See Claude Code Setup for the full guide.
Using other tools? Configure an LLM provider for on-demand analysis:
- In the dashboard, click Settings (gear icon)
- Select a provider (OpenAI, Anthropic, Gemini, Ollama, or llama.cpp for fully local inference)
- Enter your API key (not needed for Ollama or llama.cpp)
- Return to any session and click Analyze
Individual commands reference
These commands are available if you want to run steps separately:
code-insights init # Customize settings (optional — runs automatically on first use)
code-insights sync # Sync sessions without opening the dashboard
code-insights dashboard # Start the dashboard server (auto-syncs first)What's next
- The Dashboard — Tour of the embedded dashboard
- Terminal Analytics — Stats in your terminal without the dashboard
- Insights & Analysis — Deep dive into LLM-powered insights
- CLI Commands — Every command and flag