KeelenUse case — GLM (Z.ai)
GLM made agentic coding cheap. Keelen makes it shippable.
Connect a Z.ai API key to a project and use Requests to plan work. A connected project can turn Requests into planned tasks and implement changes in isolated runs. Reviewable pull requests and decisions return under the project's capacity and controls. Provider billing and allowances remain terms of the Z.ai plan.
Connect GLM with a Z.ai API key, Claude Code with a Claude subscription or API key, Codex with a ChatGPT sign in or OpenAI API key, Kimi with a Kimi Code membership key or Moonshot API key, or Grok with a SuperGrok sign in or xAI API key. Choose an engine per project. The Indie plan is $29/month: 200 iterations, 3 projects, 2 concurrent runs, and one scheduled project on the 24/7 loop.
- Connect your Z.ai key so GLM can progress your roadmap on GitHub.
- Configured controls can include adversarial review, a verified clean-checkout command, required checks, and manual merge or a review window. Red-first proof requires a verified project test command.
- Provider billing, availability, and limits remain terms of the connected Z.ai plan; Keelen's platform tier is separate.
- Switch a project between GLM, Claude Code, Codex, Kimi, and Grok while keeping its roadmap.
The volume engine
Use your connected Z.ai key for defined maintenance work: test debt, refactors, dependency updates, or small fixes. Requests become prioritized roadmap items and tasks with acceptance criteria. Eligible runs implement changes in isolated VMs and return work under your review and merge policy. Check the allowance and usage prices of the specific provider plan you connect.
Trust the gates, not the vibes
A separate adversarial review examines the diff without the development conversation. A verified project test command enables red-first proof and a clean-checkout test run. Gated auto-merge requires the configured repository checks and a review window. Review the actual results, including failures and missing evidence, before accepting the change.
A night on the loop
The scheduler picks your top task, boots an isolated VM, clones the repo, and works the task on GLM. Where a verified test command is available, it produces test evidence before and after implementation. Work follows your PR or branch policy, and decisions that need you return with the task's acceptance criteria. Check the actual progress and remaining decisions when you return.
Honest limits
Provider weather is handled for you: when Z.ai reports a usage limit or an overload window, runs back off and retry on a budget — retries never count against your iterations, and the loop resumes on its own. Ambiguous requests become clarifying questions rather than guesses. And your test suite bounds the autonomy: a repo with no tests onboards in an honest no-test posture until the loop's own PRs add one.
Recommended plan
Indie — $29 / month
FAQ
What GLM credential does Keelen need?
An API key from Z.ai (or Zhipu) — create it in your account and paste it once; it looks like an id and secret joined by a dot. The key is encrypted at rest, injected only into the single-use VM that runs your iteration, and deleted the moment you revoke it.
Is GLM good enough to run unattended?
Evaluate GLM on your repository and the checks it can actually run. Keelen supplies separate diff review; a verified test command enables red-first and clean-checkout evidence. Your configured required checks and merge policy determine when work can land. Keep manual review for decisions the available evidence cannot settle.
Is Keelen GLM-only?
No. Connect Claude Code, Codex, GLM, Kimi, or Grok with your own supported subscription or API key. Engines are per project: run GLM on the side project and Claude Code on the client repo, side by side in one workspace.
What does Keelen cost on top of my GLM plan?
A flat platform fee with zero token markup: Indie is $29/month for 200 iterations, Operator is $79/month for 1,000, and Agency is $299/month for 5,000. Your GLM usage bills to Z.ai under that provider's terms. Iteration and pull-request counts vary with the repository and request.