Blume groups coding-agent work into deliveries with revision-linked tests and screenshots for human review.
About Blume
Blume brings parallel coding-agent work into one review board. Each delivery keeps the brief, task tree, agent runs, code revision, tests and screenshots together, so you can review the evidence and approve a specific revision.
Its setup connects Codex or Claude Code through a plugin. Agents plan, build and test while Blume tracks what remains. The product preview uses illustrative data, and the homepage currently offers a waitlist.
A shared, versioned model of your product's language and business rules is on the roadmap.
The homepage currently offers a waitlist.
Pricing: The homepage currently offers a waitlist.
Features
- Parallel coding-agent delivery tracking
- Briefs, task trees and agent runs together
- Evidence pinned to the code revision
- Test steps and screenshot review
- Codex and Claude Code plugin setup
FAQs
Common questions about Blume







