KeelenAnswer · unattended work
I have no time for my side project. Can AI work on it for me?
Updated 2 October 2026
A loop can help a side project when the project has gates. Even a plausible change still needs evidence and a human decision. A better prompt is not the fix. What you need is infrastructure: a test suite, a CI pipeline, and a rule that nothing merges while a check is red. With those in place, an autonomous loop such as Keelen can work your code while you are at your day job. It opens one tested pull request at a time for you to review when you get to it.
The realistic split is this: you keep the direction and the review. You hand over the hours. That is what was actually missing, and it is the part that does not need you.
What to do, in order
Accept the real constraint
The blocker on a stalled side project is almost never ideas or skill. It is a supply of uninterrupted hours you do not have. Solve for hours. Motivation alone will not unblock the work.
Build the gates before the autonomy
Add a smoke test, tests on the main paths, and a CI workflow that runs on every pull request. Without those, you cannot check unattended work. You will end up reading everything by hand. That costs the hours you did not have.
Write your intent down where a machine can read it
A standing list of clear requests beats a chat session. It survives you not being there. Vague items produce vague work, whoever implements them.
Keep the work in small pull requests
Keep one task per pull request. Then each change is small enough to review in the ten minutes you actually have. It is also easy to undo if it is wrong.
Review on your own schedule
Set the project to open pull requests and wait. Or set it to merge only on green CI, after a review window. Either way the queue waits for you. You do not wait on the queue.
Why unsupervised agents get a bad reputation
Many people try this: point a coding agent at a repository and leave it running. It produces a large amount of code that looks plausible. Some of it is wrong, and nothing decides which is which. People then say autonomy does not work. The real lesson is that autonomy without verification does not work. That is also true of humans. It is why professional teams use code review and CI.
What the gates are
A separate adversarial reviewer reads the diff. It does not see the conversation that wrote the code. The review model comes from your project settings. Red first proof and clean checkout tests need a verified project test command. Gated auto-merge needs the repository checks and a review window you set up. Plan review, manual merge, and branch-only mode each pick where a person must decide. Check the real evidence before you accept the result.
What it costs to run overnight
The Indie plan is $29 a month. You get 3 projects, and one can run on the 24/7 schedule. Keelen runs on credentials you connect. Anthropic Claude works with a Claude Code subscription login or an API key. No Max plan is required. You can also connect OpenAI Codex, GLM, Kimi, or xAI Grok. Coding-engine token use bills to your own provider account at cost. Keelen charges one flat monthly fee and adds no markup. If you already pay for a supported AI subscription, it sits idle most of the day. This is a way to use capacity you have already bought.
When a decision needs you
Your approval settings still apply. Plan review is a separate opt in. When it is on, a human approves the plan before any code is written. With manual merge, Keelen opens the PR and you click merge. A real dead end can become a plain English decision card and pause the project.
When a provider limit has a reset time
Keelen can pause for a recognized timed session or usage limit. It uses the provider reset time when it can parse one. Work becomes eligible to resume when the pause ends. Reset handling differs by engine. Some replies give no usable reset time. Those use a bounded retry wait. That wait can grow after more usage limit failures. A retry deadline is not proof of the provider's real reset.
When a second engine can take over
A second engine can take over while the main one cools down. It must be configured and connected. Its credential must be usable. It must have no live cooldown of its own. Other scheduling controls and decisions still apply. This does not bypass provider limits. It does not lower the provider bill. It does not promise work will not stop. It does not promise a set amount of output overnight.
An optional daily or weekly email
The progress digest email is off by default. You can opt in to a daily or weekly email. It covers merged work and open PRs. It also covers paused or blocked work and decisions that need you. The email reports progress. It does not promise completed work or resolve decisions for you.
When Keelen is not the answer
- Your project has no tests and you do not want any. Then unattended work cannot be verified, and you should not run it.
- You enjoy the coding and want to keep doing it. This buys hours you may not want back.
- The project keeps changing course every week. A standing roadmap suits work with a steady goal.
Hand the work to a loop
Connect a repository and write what you want in plain language. Review the tested pull requests that come back.
FAQ
My agent stopped overnight. It was stuck on a permission prompt or a usage limit. How do I keep it running?
You may still need to approve some steps. Plan review can ask a human to approve the plan before code is written. With manual merge, you click merge. For a timed limit it knows, Keelen can pause until a parsed reset time. Work becomes eligible to resume when the pause ends. Some replies give no usable reset time. Those use a bounded retry wait. That wait can grow after repeated usage limit failures. A retry deadline is not proof of the real reset. A second engine that is configured and connected can take over while the first one cools down. Its credential must be usable. It must have no live cooldown of its own. Other scheduling controls and decisions still apply. The progress digest email is off by default. You can opt in to a daily or weekly email. It covers merged work, open PRs, and decisions that need you. None of this bypasses provider limits. None of it lowers the provider bill. None of it promises work by morning.
Will it break my project while I am asleep?
It cannot merge while a required CI check is red. Before that, your own test suite runs on a clean checkout. If you want a harder guarantee, set the project to manual merge or branch only. Then nothing reaches your main branch without you.
How much of my own time does it still need?
You write requests when you have new intent, and you review pull requests. Both fit in short sessions, which is the point. What it removes is the long block you never get.
Can it work on a project that has no tests yet?
Yes, and the first requests should be the tests and the CI workflow. Until those exist, keep the project on manual merge and review each pull request yourself.
Do I need a Claude Max plan?
No. Keelen uses whatever credentials you connect. That can be a Claude Code subscription login or API key. It can also be Codex, GLM, Kimi, or Grok. Grok accepts a SuperGrok sign in or xAI API key.