Cosine Genie
London-built agent that turns Jira and GitHub tickets into opened PRs
Overview
Genie is Cosine's autonomous coding agent: it reads your codebase, plans a fix, writes and tests the change, then opens a pull request — assigned from GitHub, Jira, Linear or Slack. Built for engineering teams that want tickets closed without babysitting every step.
Pricing
Pricing shown for reference only. These figures reflect RECATOOLS research as of 12 Jul 2026 and may be out of date or incomplete. This is not financial or purchasing advice — always confirm the current price on the provider’s official website before making any decision.
- Up to 100 projects
- Full feature access on free tasks
- 5M credits/mo
- Seat-based billing
- 60M credits/mo
- Priority support
- Custom deployment
- Dedicated support
What you can produce with Cosine Genie
- Ticket-to-PR workflow via GitHub, Jira, Linear or Slack
- Codebase-aware context retrieval before planning a fix
- Automated test running before opening a pull request
- Public CLI for local, terminal-based use
- Free tier: 80 tasks across up to 100 projects, no card required
- Enterprise cloud, VPC or air-gapped deployment options
ASEAN Perspective
Cosine Genie in Southeast Asia
ASEAN-region availability and pricing notes coming soon. Drop the editorial team a note via /contact/ if you can supply local context (Singapore/Malaysia/Indonesia/Thailand/Vietnam).
Genie built its reputation on eye-catching SWE-Bench numbers, though Cosine has never submitted results to the official leaderboard, so treat the benchmark claims as self-reported rather than independently verified. What holds up better is the workflow: assign a ticket from GitHub, Jira, Linear or Slack, and Genie retrieves the right context, writes the fix, runs tests, and opens a PR for human review rather than merging on its own.
It competes with Devin, Factory and a growing field of background coding agents, so the real test is how it performs on your own repo, not a benchmark chart. The free tier (80 tasks, no card) makes that cheap to find out. Seat-based pricing gets expensive fast on the Professional tier, and independent review-site coverage is thin, so lean on a hands-on trial over marketing copy.
What people say
Cosine is a London-based team (YC-backed) whose pitch centers on Genie, a fine-tuned coding model trained on real software-engineer workflows rather than generic code completion. The company's own benchmark claims are aggressive — a self-reported 30.08% on SWE-Bench and 43-44% on SWE-Bench Verified, which Cosine says beat Amazon Q and Cognition's Devin at the time. Hacker News commenters flagged that these numbers don't appear on the official SWE-Bench leaderboard, because Cosine declined to publish the methodology needed to qualify — worth knowing before treating the scores as apples-to-apples.
Coverage from VentureBeat and the Latent Space podcast (an interview with co-founder Alistair Pullen) describes Genie's differentiator as training data: examples of human engineers actually working, not just finished diffs. In practice, that shows up as a ticket-to-PR workflow — pull a task from GitHub, Jira, Linear or Slack, and Genie retrieves codebase context, drafts a plan, writes and tests the change, then opens a pull request for review rather than auto-merging.
Independent user reviews are scarce; Genie doesn't have a meaningful G2 or Capterra footprint as of mid-2026, so most of what's public is vendor blog posts, benchmark write-ups and YC/Product Hunt listings rather than at-scale user feedback. The company does maintain a public CLI repository on GitHub with regular releases (54 releases as of July 2026), which at least gives outside developers something concrete to inspect beyond the marketing site.
Pricing is seat-based: a free tier covers up to 80 tasks across 100 projects with no card required, Hobby runs $20 per seat monthly with 5 million Cosine Credits, and Professional jumps to $200 per seat monthly for 60 million credits. Enterprise deployments (cloud, VPC or fully air-gapped) are quoted separately. That's a steep jump from Hobby to Professional for teams that outgrow the entry tier, and credit consumption in agentic coding tools tends to scale with iteration count rather than task complexity, so budgeting ahead of a team rollout is worth doing carefully.
Summary of public user & expert reviews, compiled by RECATOOLS.
About this listing
This entry was compiled from publicly available data including Cosine Genie's official website, press releases, documentation, and reputable third-party publications. RECATOOLS is not affiliated with Cosine Genie unless explicitly stated.
Third-party AI tools update their pricing, features, availability, and policies frequently. Information here may be outdated by the time you read this — we make reasonable efforts to keep listings current, but cannot guarantee absolute accuracy.
For the latest details, please refer to Cosine Genie directly →
Spotted something out of date? Suggest an update →
More in Code & Dev Tools
