You deserve better CI. Cursor, OpenAI, Anthropic, Perplexity, vLLM, Uber, Airbnb, Canva, Planetscale and more standardized orchestration with Buildkite.
Trial? More like Try All! 30 days, everything unlocked, zero commitment. No credit card. No feature gates. A real human engineer on standby.
After 30 days, you choose your adventure, without surprise invoices or auto-upgrades. Your decision.
Play with Buildkite → buildkite.com/pricing
2/ At Elastic, Claude Code now runs inside Buildkite CI. When a dependency-update PR breaks the build, the agent reads the logs, iterates edit-compile-test, and commits a fix.
First month: 24 broken PRs fixed. About 20 days of engineering work saved.
We’ve seen workloads double in the past three months on @buildkite. Will be well on track to 10x at this pace on an already huge volume base. We run CI for the fastest moving companies like Shopify, Uber and most of the frontier AI labs with a similar workload profile in terms of stateful guarantees. The amount of code being pushed is truly unprecedented.
Schrödinger's merge wasn't on the downtime bingo card.
You deserve better CI. And a status page that hasn't been busy all week.
Our new trial unlocks everything in the product with zero commitment. And, we promise the merges will stay merged.
Start building here→lnkd.in/gAu5bfbA
SF friends 👋
We’re hosting our first Buildkite Localhost, a casual meetup for folks interested in developer tooling.
No talks, no pitches. Just food, drinks, and real conversations about AI-generated code, scaling CI, and keeping pipelines sane.
Perry’s Embarcadero
Jan 28 from 6–8 PM
luma.com/4a8vm2zb
Building an MCP server for Buildkite taught us three important lessons about building great agent tools:
- Wrapping your REST API isn't enough — agents need structured, clean data
- You only control tool shape, not agent strategy
- LLMs are lazy — make the right path the easiest one
See how we applied this to CI log navigation: buildkite.com/resources/blog…
Congrats to all the LEGO winners who stopped by the booth @ #AWSreInvent (check the 🧵)
Can't wait to see these builds once they're complete! 🧱
P.S. If you're in Vegas & haven't come to see us yet, we might have a few sets left 🤞🤞🤞
To get help from an LLM with a broken CI build, you have to give it a log — but build logs can be huge and filled with noise that nave nothing to do with what actually went wrong.
In this post, we share how we solved this problem in the @buildkite MCP server by asking the LLM to critique its own work: buildkite.com/resources/blog…