Skip to content
IndieGent
Why IndieGentmacOS · $19 once

The math changed.

One person can only do one thing at a time. That is the constraint behind every stalled project you have ever owned, and it has just moved. IndieGent is what I built on the other side of it.

01The old math

It was a fantasy or a burnout plan.

Everyone working alone keeps the same list. Four or five projects. One of them shipping, two of them stalled, one that would take a weekend if a weekend ever came free. The work is rarely hard. There is just one of you.

The two ways out were both bad. Hire, and you take on payroll, management and a monthly number before there is revenue to cover it. Don’t hire, and you work down the list one item at a time while it grows from the other end.

I picked the second one and called it focus.

02What changed

Agents changed the arithmetic.

A coding agent can take an issue, read the repository, write the code, run the tests and open a pull request. That stopped being a demo some time ago. It is a Tuesday.

But one agent in one terminal is only a faster you. The multiplier was never the agent. It is how many you can have working at once before directing them costs more than doing the work yourself.

That ceiling is the real problem, and it is made of admin. Every agent you add wants a branch, a clean directory, a window to watch and an answer when it gets stuck. Six of those is not a team. It is a second job.

03The point

Direction should cost seconds.

So that is the part I made cheap. Tick an issue and send it. The workspace is cut and dressed for you, the branch is right, and the session can run with no terminal and no window at all. What comes back is a pull request that closes the issue you ticked, and nothing left behind on your Mac.

That sounds like convenience. It isn’t. When dispatching work costs seconds you stop rationing it — you send six things instead of one, and spend the morning deciding which of them deserved to exist.

You spend your time deciding. Agents spend theirs doing. Every other decision in this product is downstream of that one sentence.

one ordinary morning

09:02you · three issues ticked · three sessions dispatched
09:04acme-api #86 · session opened · workspace cut · branch dressed
09:31PR #91 opened — closes #86acme-api · +182 −44 · tests green
10:02PR #132 opened — closes #131papertrail · wrapped up · workspace deleted
11:26PR #49 opened — closes #47castaway · headless · no window ever opened
Twenty minutes of direction. Three pull requests. You were on a call for most of it.
04What I sell

Time, and nothing else.

Every tool either makes you money or saves you time. This one returns time to people who convert time into product, which makes it both.

And the hour that comes back is not an hour of typing you avoided. It is an hour of judgment: what to build, who to talk to, what to kill. Agents are not better than you at that part. They should be nowhere near it.

Nineteen dollars, once. Work out what an hour of yours is worth and the sum stops being interesting.

05Who it’s for

A workbench, not a platform.

An indie developer with more repositories than evenings. Two people who want to ship like six. Anyone already running an agent CLI and tired of babysitting it.

Not platform teams. Not an org chart. There is no seat count, no admin console and nothing to roll out on Monday. It is a workbench, and it belongs to the person doing the work.

It runs on your Mac, against your GitHub, driving the agent CLI you already pay for. I am not in the middle of your code, and I would rather not be.

The standard

  • Your attention is the scarce resource. Not compute.
  • Work should happen whether or not you are watching.
  • A tool that needs managing is not help.
  • Every session ends clean, or it does not count.
  • The person still decides.

The list hasn’t moved.

Take one thing off it tonight and send an agent at it. Seven days free, no card. Then nineteen dollars once — with the agent CLI you already use.