Codex moved their CI off GitHub Actions, expected to pay back in months
Buildkite wants GitHub webhooks for every repo, the kind of setup that's tediously annoying by hand. Codex built ephemeral runners on refurbished Mac minis and migrated each repo. Payback: two to three months.
Done withCodexSpotted· by HN user ElFitz· Oct 18, 2026
Result Their colleagues are now considering it too
Spotted: found in the wild. Details are as reported by the source above and have not been independently verified.
What they did
They had Codex set up ephemeral VM runners on a couple of low-spec refurbished Mac minis, wire up Buildkite’s API with a short-lived API token, and migrate their repositories to it one by one, after deciding the GitHub webhook setup that Buildkite requires for each repo would have been tediously annoying to do by hand.
How it went
The migration went well enough that they expect it to pay for itself within two to three months, and after an outage that day, they suggested their colleagues try the same setup, which they’re now considering.
Try it yourself with Codex
Have [your coding agent] set up a full CI migration from [your current CI provider] to [a new provider, e.g. self-hosted runners on spare hardware], including provisioning ephemeral runners, wiring up API tokens, and migrating your repositories one at a time. Test each migrated repo before moving to the next.
A starting prompt written by Agents in the Wild. Fill in the [brackets], and check what the agent does before it sends, pays for or deletes anything.
Building their first AWS project by vibe coding, the writer designed the architecture and let Codex write the code. A flawed first design took a day, against an estimated three or four; the redesign also took a day and scaled.
Four years after shelving a homemade RSS reader and selling off its $500 domain, the writer built it this year with Codex and Claude. They use it daily, along with 30 other strangers who found it public.
Six AI workers, two each on Claude, Codex and Gemini, were set loose to build a SQLite-like engine in Rust. It passed 64 queries of a 1,000+ query suite, with slop like redundant buffer copies and no concurrency.
They already run local reviews with Codex and Claude Code. Assigned as a reviewer in GitHub, Copilot finishes in a couple of minutes and raises valid issues they'd never found.
Code from one AI is hard to trust blind. This developer runs a Docker harness where Claude writes and Codex reviews, works in it most of the day, and open-sourced it as moarcode.
Optimizing a Taichi kernel had each agent looping on its own. Putting Claude Code and Codex in one discussion to combine their ideas cracked it, saving several hours.