Track Codex adoption, impact, and cost in Swarmia

Codex is now one of Swarmia’s built-in AI tool integrations, alongside GitHub Copilot, Cursor, Claude Code, and CodeRabbit. Connect your ChatGPT Enterprise workspace to bring Codex data into Swarmia, including:

  • Adoption. See who’s using Codex each week and how that spreads across your teams.
  • Detection. Swarmia flags the pull requests Codex worked on, whether that’s local changes from the CLI or IDE, cloud agent pull requests opened by the Codex connector, or reviews left by Codex.
  • Impact. See how Codex-assisted pull requests compare on throughput, cycle time, batch size, and time to review.
  • Cost. See what Codex tokens cost per person and per team, next to your other AI tools. In the focus summary, that spend is attributed to the work it went toward.

Together, these views show whether Codex is still being experimented with or has become part of teams’ day-to-day workflows. You can also see where adoption is strongest, how Codex-assisted work performs, and how much pull request activity Codex contributes to.

How to get started

Go to settings → AI tools → Codex and follow our guide on connecting Codex.

You’ll need a ChatGPT Enterprise admin to create an admin key for your workspace. Cloud agent pull requests and Codex reviews are detected from your existing data, so they don’t require any additional setup.

The Codex Enterprise Analytics API is only available on ChatGPT Enterprise. If you’re on a Business workspace, email us at hello@swarmia.com and we’ll add you to the waitlist.


Subscribe to our newsletter

Get the latest product updates and #goodreads delivered to your inbox once a month.

More changelog updates
· Sep 14, 2026

Set targets for your key metrics

Key metrics show you where your engineering organization stands today. Now you can also show where you want it to go. You can set a target at the organization level, and it applies to all of your teams. From there, you can adjust the target for individual teams where it makes sense. If an…
Read more
· Sep 11, 2026

Find out how much of your cycle time is spent in draft

Cycle time is an aggregate metric, and alone it doesn’t describe where things slow down. Now you can look at each pull request stage and analyze outliers in each — whether it’s review time, merge time, in progress time or time as a draft. Draft time shows how much of a pull request’s cycle time it…
Read more