Who is building/using agents for Linear?
Linear for Agents is now GA. Who is building or using an agent for Linear?
We recently built one for using Bucket feature flags within Linear (and shared some of our lessons learned). Would love to hear from others building or using agents. We're just getting started with our Linear integration and are keen to get feedback. We've plans to add:
Feedback/adoption result summaries to the agent
Support for an access summary embed whenever a link to a feature on Bucket is dropped
Plus, integrate our feedback function with Linear's customer requests - making it possible to capture customer requests directly from within your application
If you've tried it, or if you're up beta testing what's next, let me know.
Replies
I’ve built a small internal agent that connects Linear with our knowledge base and Slack. The feedback summary idea you mentioned is 🔥, feels like a natural next step to tie user feedback into decision loops.
Bucket
@aayaz_gul Thanks Aayaz, I think you're right. Feedback can often be too unwieldy to get back into decision loops, so we're trying to integrate it into the dev workflow.
We’re testing an internal agents that auto-updates issue statuses based on GitHub activity. It’s not perfect, but it reduce a lot of manual updates.
We’re trying a very lightweight assistant that posts daily summaries of Linear changes to our team Slack. Nice little boost for async updates.
Product Hunt
@md_ansar_ray I like that idea! It can be tough to keep up with everything that gets updated/completed every day in @Linear
As a member of a team developing a knowledge assistant for development teams, I recently combined it with Linear to highlight the context of tickets from documents and previous projects. The way your agent pulls information into Linear based on feature utilization is fantastic. Connecting consumer mood and usage to problems directly can be revolutionary, particularly as teams accelerate and ship more iteratively. I completely see your approach opening the door to better product choices.
Bucket
@anish_hessa That's the hope. I think where it can be most impactful is to get product feedback back to devs directly, so they can be more autonomous and iterate faster.
We build an agent that listens to our sentry error stream and auto creates linear issues for new crash signatures.
Bucket
@steven_granata ooh, that sounds handy
Warestack
We’ve built both one-shot and chat-based agents for Linear. Our logic interprets GitHub events such as pull requests and deployment reviews, and propagate context back to related Linear tickets. In our assistant, users can also ask questions about their Linear workspace (from issue status and ownership to recent activity) and get contextual answers in real time.
Bucket
@dkargatzis Oh nice, so it's updating issue statuses based on PRs being merged and reviews being completed?
Oh! We use an app called Alfred.sh - it's an AI PM Agent. I use it all the time. I have a conversation with Alfred over Slack, and when I'm ready I tell Alfred to write a ticket to Linear. It's great. I can make edits to tickets. I can ask him to check if a ticket has already been made. I've asked the CEO to join Product Hunt so he can join this conversation! :D
@laura_cruickshanks Thank you Laura! Sorry just joined.
Yes, we have been very early adopter of @Linear agent! Quick takeaways:
https://www.alfred.sh/ is a really a collaborative PM teammate that human PMs, engineers, etc can delegate tasks to. This means Alfred has been writing Linear tickets for Human PMs, engineers, etc.
Working well:
"on behalf of" feature -- this is awesome. It allows Alfred to attribute the work to humans (and not take credit for their work)
For example Alfred takes a rough PRD, or ideas and massively created 10 fully fleshed out Linear tickets, the on-behalf-of field shows that the human PM owns those tickets! This makes human PMs looks super awesome while the agent did the heavy-lifting of writing up each individual tickets with all the details (acceptance criteria, tickets, etc) filled in.
the api is also pretty awesome for then attaching docs, images, etc.
Not fully baked:
The api doesn't work well yet when it comes to attaching videos. Happy to share more! Please feel free to reach out!
Haye
My simplest idea for Agents is that they should be able to summarize what I've done in the past. Right now, it seems like I have to manually organize everything, which is too troublesome
Bucket
@hi_caicai Like a daily digest of things you did?
Haye
The most crucial thing isn't about my own affairs, but rather what the research and development staff and the entire team have done over the past period
Relevance AI, and individual developers are building AI agents for Linear to automate tasks like issue creation, project updates, and workflow management.
The ability to pull customer feedback right into Linear would be a game-changer for us. Keen to follow your progress.