Skip to content
| Marketplace
Sign in
Visual Studio Code>Other>TinyAI ObservabilityNew to Visual Studio Code? Get it now.
TinyAI Observability

TinyAI Observability

tinyai

| (0) | Free
Native VS Code extension for Copilot-era TinyAI coding telemetry.
Installation
Launch VS Code Quick Open (Ctrl+P), paste the following command, and press enter.
Copied to clipboard
More Info

TinyAI Observability

Private VS Code extension for collecting Copilot-era coding task telemetry.

Usage

  1. Install the private .vsix.
  2. Reload VS Code.
  3. Keep using Copilot Chat / Agent mode normally.
  4. The extension automatically scans local VS Code Copilot transcript files and uploads full user/assistant messages when tinyaiObservability.captureConversationText is enabled.
  5. In Agent mode, enable or allow the tinyai_specs tool. Copilot can then call the tool directly when the prompt needs personal specs.
  6. For sticky TinyAI chat, start a Copilot Chat thread with @tinyai; the participant is sticky, so follow-up turns in that thread do not need @tinyai again.
  7. Open the left-side TinyAI activity bar item for task controls: Capture Copilot Local Transcripts, Start Task, Record Feedback, Record Adoption Snapshot, End Task, and fallback transcript import.

The extension records task lifecycle events, spec/catalog file access, code change summaries, full conversation text for TinyAI-owned chat surfaces, local Copilot transcript JSONL files, clipboard/editor-imported Copilot transcripts, feedback, adoption snapshots, and upload retries.

Regular Copilot Chat is captured from local VS Code workspace transcript files:

$HOME/Library/Application Support/Code/User/workspaceStorage/<hash>/GitHub.copilot-chat/transcripts/*.jsonl
$HOME/Library/Application Support/Code/User/workspaceStorage/<hash>/chatSessions/*.jsonl

Those records include user and assistant message text when VS Code persists them. They are marked derived because the extension reads local persisted files rather than receiving a first-party Copilot Chat event. Direct capture still comes from the sticky @tinyai participant and the tinyai_specs Agent-mode tool. Inline completions are not per-suggestion exact; code adoption is estimated from task-window diffs and retention snapshots.

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