Cognition Launches Devin Bots, AI Managers That Lead Teams of Coding Agents
Cognition launched Devin Bots on Nov 17: persistent agents that use memory to lead teams of Devin coding agents on engineering projects lasting days to weeks.

What happened
Cognition introduced Devin Bots in its official blog as persistent agents: given a goal, a Bot breaks it into pieces, hands the work to multiple Devin sessions running in parallel, evaluates the results and decides what to pursue next, keeping context throughout; only decisions needing human judgment go to a Needs You inbox for one-click approval. Users work with Bots over Slack, iMessage or the Devin app, and multiple Bots can coordinate with each other.
Cognition says Bots pick different models per stage to balance speed and cost; each Devin runs on its own cloud machine, code goes through the same tests as human-written code, and teams can configure low-risk changes for automatic approval and merging. Bots ship with six starting points: Software Factory, Android Engineer, Chief of Staff, Latency Optimizer, Product Engineer and QA Engineer.
Cognition offered three figures of its own: testers merge 4.5 times more pull requests per message on average; one game studio's weekly merged pull requests rose from about 50 to about 1,900; and an Android Bot built Devin's Android app in 44 days, merging 1,134 pull requests, writing 344,651 lines of Kotlin and closing 1,217 tickets. Early customers include Decagon, Fireworks AI and Reducto, and Bot usage grew tenfold within two weeks.
Key facts
- Launch: Nov 17; available at devin.ai/bots; positioned as persistent manager agents.
- How it works: takes a goal, splits the work, evaluates results and keeps context; only decisions needing human judgment go for approval.
- Channels: Slack, iMessage and the Devin app; shareable with a whole team; multiple Bots can coordinate.
- Model strategy: models picked per stage; Cognition's own models focus on software engineering and code review.
- Company figures: testers merge 4.5x more pull requests per message; a game studio's weekly pull requests rose from about 50 to about 1,900; an Android Bot merged 1,134 pull requests and wrote 344,651 lines of Kotlin in 44 days.
- Market context: Cognition is valued at $48 billion with a $1 billion annualized revenue run rate (Fortune); OpenAI shipped Dots in late September and Anthropic shipped Duet.
Why it matters
Coding agents move from taking tickets to managing projects, and Cognition points to testers merging 4.5 times more pull requests per message; the figures are the company's own, with no independent verification.
Follow HIMIAO Horizon
No sign-up · Unsubscribe any timeNotifications
A ping on your phone or computer when the brief is out. No email needed.
Comments
Related
Today
Nov 19 Thursday- Brief
- GitHubGitHub Weekly Vol.6: 7 Projects Worth a Look This Week
Tested or read — weeklyAll →
- Quiz
- Call
Latest news
All →- Nov 19MAI Code 2 vs Claude Opus 5.5 and GPT-6.1 Sol: $0.50 Input vs $4, Promo Price Until May 2027
- Nov 18Study: Frontier Models Can Be Tricked Into Marking Civilian Targets
- Nov 18Nvidia Reaffirms Open-Source Strategy, Hugging Face Deal on Track
- Nov 17Klarna Launches AI Assistant That Shops, Books and Runs Errands
- Nov 18Palo Alto to Buy Applied Compute for $4.5 Billion in All-Stock Deal
- Nov 18NGINX2Root Exposed: Five 9.2-Severity Flaws Let Attackers Run Code on Nginx Without Logging In
™