Skip to content
| Marketplace
Sign in
Visual Studio Code>AI>Copilot Insights (Local)New to Visual Studio Code? Get it now.
Copilot Insights (Local)

Copilot Insights (Local)

Rajat

|
2 installs
| (0) | Free
Automatically indexes your Copilot chat history into a private local SQLite database so you can search past sessions and analyze token usage and cost — all stored on your machine.
Installation
Launch VS Code Quick Open (Ctrl+P), paste the following command, and press enter.
Copied to clipboard
More Info

Copilot Insights (Local)

A 100% local VS Code extension that indexes your GitHub Copilot chat history into a private SQLite database so you can search past sessions, recover context, and track token usage and estimated cost - all on your machine. Nothing is ever uploaded your data never leaves your computer.

Optional: connect provider usage APIs (OpenAI, Anthropic, Mistral, Kimi) to see your direct API token usage alongside Copilot, in one dashboard.

Features

  • 100% local & private - everything is stored in a SQLite file in the extension's storage; no telemetry, no servers.
  • Automatic capture - a background watcher scans VS Code / Cursor chat session storage and indexes new or changed sessions incrementally.
  • Full-text search - FTS5 over every message, from the command palette or @insight /search.
  • Exact tokens & billed credits - reads each turn's exact prompt/output tokens and the billed premium credits straight from the session logs.
  • Analytics dashboard - a tabbed webview (This Workspace, All Sessions, Overview) with stat cards, a GitHub-style token activity calendar, per-model breakdown, and by-workspace and by-provider breakdowns with per-session estimated cost.
  • Browse & recall sessions - sessions grouped by workspace; click any row to read the full conversation, tag it, or export it.
  • Multi-provider usage (optional) - pluggable connectors pull daily token usage from OpenAI, Anthropic, Mistral and Moonshot/Kimi usage APIs. Keys are stored in VS Code SecretStorage, never in settings or the database.
  • Session tags & notes - annotate any session; tags surface in the dashboard.
  • Export - one session as JSON, all sessions as CSV, or a Markdown report.
  • Status-bar cost - today's estimated spend + billed credits, with an optional daily budget.
  • Copilot integration - @insight chat participant (/search, /sql, /stats, /context), plus language-model tools so Copilot can search your history on its own (#insightRecall).

Getting started

  1. Install the extension (from the Marketplace, or via the Extensions view → … menu → Install from VSIX…).

  2. Open the dashboard — either click the Copilot Insights icon in the Activity Bar (the vertical bar on the far left), or use the Command Palette:

    • Open the Command Palette: Cmd+Shift+P (macOS) or Ctrl+Shift+P (Windows / Linux).
    • Type Insight and choose Insight: Open Dashboard.

    Your existing Copilot history is indexed automatically in the background.

  3. Explore the Overview tab for totals, the token activity calendar, and the per-model / per-provider breakdowns. Use All Sessions to search and read past conversations.

Connect a provider (optional)

To include your direct API usage (outside Copilot):

  1. Open the Command Palette (Cmd+Shift+P / Ctrl+Shift+P) and run Insight: Connect Provider, then pick a provider from the list.
  2. Paste an API key when prompted (for OpenAI/Anthropic this is an admin usage key). Keys are stored securely in VS Code SecretStorage — never in settings or the database.

Usage syncs on startup and on a schedule. To refresh immediately, run Insight: Sync Provider Usage Now; to remove a provider and its imported data, run Insight: Disconnect Provider.

Commands

Run any of these from the Command Palette — open it with Cmd+Shift+P (macOS) or Ctrl+Shift+P (Windows / Linux), then type Insight and pick a command:

Command What it does
Insight: Open Dashboard Opens the analytics dashboard in the editor
Insight: Search Past Sessions Full-text search across all indexed sessions
Insight: Reindex All Sessions Force a full re-parse of every session file
Insight: Tag / Annotate a Session Add tags and a note to a session
Insight: Export Session as JSON Export a single session
Insight: Export All Sessions as CSV Export a usage spreadsheet
Insight: Export Report as Markdown Export a summary report
Insight: Connect / Disconnect Provider Manage provider usage connectors
Insight: Sync Provider Usage Now Pull the latest provider usage
Insight: Clear All Indexed Sessions Wipe the local index (source files untouched)

Settings

  • copilotInsights.autoIndex - index sessions automatically (default true).
  • copilotInsights.databasePath - custom SQLite file path.
  • copilotInsights.dailyBudget - daily USD cap shown on the dashboard.
  • copilotInsights.pricingUrl - model pricing table URL (fetched daily, cached).
  • copilotInsights.providers.enabled - provider connectors to sync.
  • copilotInsights.providers.lookbackDays - days of provider usage to fetch.
  • copilotInsights.providers.syncIntervalHours - auto-sync interval (0 disables).
  • copilotInsights.providers.moonshotBaseUrl - Moonshot/Kimi API base URL.

Where to change settings

Open Settings in any of these ways:

  • Keyboard: Cmd+, (macOS) or Ctrl+, (Windows / Linux).
  • Menu: Code → Settings → Settings (macOS) or File → Preferences → Settings (Windows / Linux).
  • Command Palette (Cmd+Shift+P / Ctrl+Shift+P) → Preferences: Open Settings (UI).

Then search for “Copilot Insights” to edit the options above with checkboxes and fields.

Prefer raw JSON? Run Preferences: Open User Settings (JSON) from the Command Palette and add, for example:

{
  "copilotInsights.autoIndex": true,
  "copilotInsights.dailyBudget": 20,
  "copilotInsights.providers.enabled": ["openai", "anthropic"],
  "copilotInsights.providers.lookbackDays": 60
}

API keys are not stored in settings - add them with Insight: Connect Provider (kept in SecretStorage).

How cost is estimated

Dollar amounts are computed at read time from token counts and a model pricing table (a bundled snapshot, refreshed from GitHub's published pricing when reachable). Models without a known price show - rather than a misleading $0. Anthropic models use their cache_write rate for non-cached input to mirror how prompt caching is billed. Cost is an estimate and can differ from a provider's actual invoice.

Privacy

Everything runs locally. Indexed sessions and usage live in a SQLite file in the extension's storage. Provider API keys are kept in VS Code SecretStorage. The only network calls are: (1) Fetching the public model pricing table, and (2) Only if you connect a provider - that provider's usage API using your key.

  • Contact us
  • Jobs
  • Privacy
  • Manage cookies
  • Terms of use
  • Trademarks
© 2026 Microsoft