Skip to content
| Marketplace
Sign in
Visual Studio Code>Testing>Codela — Profile AI code before it breaks prodNew to Visual Studio Code? Get it now.
Codela — Profile AI code before it breaks prod

Codela — Profile AI code before it breaks prod

Codela

|
7 installs
| (0) | Free
Profiles your local test runs, cross-references them against your team's production incident history, and flags failure patterns inline in your editor.
Installation
Launch VS Code Quick Open (Ctrl+P), paste the following command, and press enter.
Copied to clipboard
More Info

Codela — VS Code / Cursor Extension

Turns traces and profiles — from local test runs and your production APM — into AI agent memory. Profiles your test runs, looks up the function you're editing in Datadog, and flags failure patterns inline before the PR opens.

How it works

From your test runs

  1. You run pytest / jest / vitest in VS Code (test explorer or terminal task).
  2. Codela wraps the run with OpenTelemetry auto-instrumentation, pointing at a collector embedded in the extension (localhost:4318).
  3. When the run finishes, Codela POSTs the trace to the Codela server (/profile endpoint).
  4. The server matches trace signals — query counts, span trees, call durations — against your team's production pattern library.
  5. Any matches land back in the editor as inline red/orange annotations on the offending lines. Hovering shows the pattern name, prod incident evidence, and suggested fix.

From your APM (Datadog)

If you connect Datadog, Codela uses the function you're working on as a join key into your production traces (via the @code.function tag set by Datadog's Source Code Integration). You get p50/p95/p99, error rate, and slow exemplars surfaced directly in the IDE — no manual context switching to your APM dashboard.

Credentials are stored in VS Code's encrypted SecretStorage and forwarded per-request; they never persist on Codela's servers.

Install

# VS Code
code --install-extension codela.codela

# Cursor
cursor --install-extension codela.codela

Commands

Command What it does
Codela: Profile Current File Run the profiler explicitly on the focused file
Codela: Show Production Evidence Open the panel with the matched pattern's evidence
Codela: Show Trace Waterfall Open the call-tree waterfall view
Codela: Reset Setup Clear cached setup state; re-checks dependencies on next run
Codela: Configure Datadog Credentials Set API key, App key, site, service tag — credentials go to SecretStorage
Codela: Test Datadog Connection Smoke-test the live /datadog/function query for a function you name

Development

cd ide/vscode
npm install
npm run compile
# Open this directory in VS Code and press F5 to launch an Extension Development Host

Configuration

Setting Default Description
codela.serverUrl https://api.usecodela.com Codela server base URL
codela.companyId "" Scope pattern matches to your company
codela.autoProfileOnTestRun true Auto-wrap pytest/jest runs with OTEL
codela.profileTimeoutSeconds 60 Max seconds Scalene can profile before being interrupted
codela.datadog.site datadoghq.com Datadog site (datadoghq.com, datadoghq.eu, us3.datadoghq.com, us5.datadoghq.com, ddog-gov.com)
codela.datadog.service "" Datadog service tag for this repo (e.g. payments-api)
codela.datadog.env production Datadog environment tag to query

Datadog API key and App Key are not stored in settings.json — they live in VS Code's encrypted SecretStorage. Set them via the Codela: Configure Datadog Credentials command.

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