Introducing coding sessions. Linear Agent can now triage issues, investigate the cause, write the fix, open a PR, and bring the code back for review. All shared with your team in Linear.

Linear Agent uses Claude Code and Codex to start, with more coming soon. See how we're closing the loop for teams and agents.

What is this place you're speaking at and how do I come visit

beautiful work!

@petergyang This is so cool 🤩

Can i use ny existing claude and codex subs?

No, coding sessions require AI credits from Linear to run. More on those here:

This is sickkkkkk Pair this with @MomenticAI >>> automatically classify any tests that are out of sync, batch them by root cause, fix your entire test suite with no manual intervention.

beautiful cinematography, very Linear!

what kind of announcement video style is this? I can’t even listen to what he saying. I am expecting John Wick to jump out of nowhere soon

i feel like this would make more sense if you had more shadows, more black background stuff, more like circles maybe?

Linear's position continues to improve. Rapidly approaching "Linear manages the entire lifecycle of your software development, from ticket through deployment"

amazing improvements! So cool to see the Linear pipeline coming together

Why do I need @linear for this. I can already do this today? What’s different other than not being able to use my existing LLM subs?

FYI. less face time and more ui-time in these videos. I get everyone wants their victory lap moment and the whole lets see how different roles use the same thing pitch. but... the ui is king mkay.... a close up of some dude talking about UI and not showing wtf he's doing is pointless.

Integrating the coding agent directly into the issue tracker closes the loop between bug report and deployed fix. The next bottleneck becomes review quality rather than implementation speed.

most tasteful presentation of this kind of feature I've seen, nice work

How can I use this for native iOS development? What environment does it run in? Is there the equivalent of a self hosted GitHub actions runner?

Linear is coming for it all

this looks awesome very excited to give this a try! :)

no. just be a fucking linear.

The important part is bringing the work back into the team system for review. Agent coding without that return path becomes parallel shadow work.

can you send a ✉️?

hey diem me right now

@elonmusk just buy them and merge them with @cursor_ai

So geeky. So beautiful. Gotta try Linear…

hey diem me plz

Linear Agent coding sessions: triages, fixes, PRs, review all in Linear with your team agents that actually ship code.

the real product here isn’t coding sessions it’s continuous delegation tickets stop being tasks and start being conversations between systems

the part i care about isn't the PR, it's what it does when it can't finish. fail loud with a rerun command, or quietly hand you a half-fix? i won't let an agent merge on the pass it coded. review stays the gate 🫡

the design choice that matters: "all shared with your team in linear." most ai coding tools (cursor, claude, copilot) work in private. team-visible agent work changes the audit dynamic. project managers can see what got fixed and by which agent. governance built-in.

🤯

If you're juggling multiple isolated Codex accounts across your desktop or VS Code, I built a lightweight, open-source switcher to handle it without the constant login hassle:

hey are you interested in this new AI tool?

yeah you should. everyone there actually ships. totally different vibe from other conferences. you coming?

Can Claude access figma mcp in this workflow?

You can connect MCP servers to the Linear agent here;