Try Devin for pull requests on a company codebase
A commenter's team tried Devin but found PRs slow and still needing review, so it felt like a junior developer.
Done withDevin
- What they did
- One commenter said their team at a decently sized tech company tried Devin on its codebase to get pull requests submitted. In practice, Devin produced PRs slowly, and the team still had to review the code themselves, so the human work stayed in the loop and the tool mainly acted like a very junior developer.
- How it went
- They stopped using it because the turnaround was too slow and the value did not beat ordinary review overhead. The commenter said it felt comparable to hiring a junior developer, and the pricing also was not competitive.
- Worth knowing
- The key catch was cost versus speed: the commenter said Devin was slow to submit PRs and still required full code review, so the economics did not work for them.
Try it yourself with Devin
In [repo name], take this ticket: [ticket description or link]. Make the change, run the existing tests, and open a pull request with a clear description of what you changed and why. Show me the plan and the diff before you push, and list anything you were unsure about so I can review it quickly.
Want help getting this working? Join the Vibe Coding Academy ↗ Run by Alex Finn, who also runs this site.
2 more accounts of this
Source: Reddit · Jan 1, 2025
Related use cases
Hand narrow code changes to Devin for pull requests
A team used Devin for a month on small fixes like chart formatting and API edge cases, with tests and PRs; a complex full-stack task went poorly.
Explore code to write technical documentation
A technical writer uses Devin to explore code before documenting it, plus a Slack writing agent built on Devin, reporting about 20,000 lines of docs per month.
Test features on screen and merge fixed PRs
Ryan Carson uses Devin's Test App feature to record and narrate feature clicks, fix bugs, and merge PRs with a !land command.
Build a C compiler with a team of parallel agents
Anthropic's engineering write-up on tasking Opus 4.6 agent teams with building a C compiler.
Triage failed GitHub Actions runs and self-heal flaky tests
A QA engineer's OpenClaw agent reads failed CI logs, classifies failures, posts analysis to GitHub and Slack and opens issues; flakiness fell from 8% to 0.8%, though it once wrongly altered an assertion.
Crawl documentation into clean Markdown for AI agents
The author built a documentation crawler for their AI agents that pulled 60 pages of clean Markdown in 49 seconds.
Custom agent
Get 5 like this every morning
The best things people got an AI agent to do, each with the prompt to try it.