Context

AwenesOS began as a way to stop doing agent coordination by hand. I was running coding agents across several projects, then reviewing, re-prompting, and updating a task tracker and a CRM to keep my records in line. The first version automated that for me, exactly as I worked.

From a command line to a desktop app

Decision structure

Should AwenesOS stay a command-line tool or become a desktop app?

Context
Agent work across several projects at once, where most of the effort is reviewing plans and changes and deciding what happens next.
Constraints
  • Several projects and tasks in flight at once
  • Plans and diffs need careful review
  • Approvals should be quick to find and act on
  • Other developers should be able to pick it up

Options

Keep the CLI
Fit
Quick for me to build on, and already worked for my own use.
Cost
Reviewing plans and diffs across projects in a terminal was slow, and the tool was hard for anyone else to learn.
Move to a desktop GUI
Fit
Every task’s state, and what is waiting on a decision, is visible at once, and review happens next to the work.
Cost
A larger application to build, test, and maintain.

Tradeoffs accepted

  • Accept a bigger application in exchange for review that is easier to do well.
  • Keep a command-line entry point for scripted tasks rather than throwing it away.

Recommendation

Build the desktop app around the review loop, because review was where most of the time and risk sat.

When I would choose differently

If most use were unattended automation rather than review, I would keep the command line as the main interface.

Removing the CRM integration

Decision structure

Should AwenesOS keep updating a CRM as part of the agent workflow?

Context
The CRM step existed because of how I personally track client work, not because agent orchestration needs it.
Constraints
  • Keep the core focused on agent tasks
  • Avoid integration work unrelated to agents
  • Fit different developers’ setups

Options

Keep the CRM integration
Fit
Matched my own workflow end to end.
Cost
Added complexity and tied the product to one person’s way of working.
Remove it
Fit
Keeps the product about flexible agent-task management.
Cost
I lost an automation that was convenient for me.

Tradeoffs accepted

  • Give up a personal convenience so the product works for people with different tools.

Recommendation

Remove the CRM coupling and focus on flexible agent-task management.

When I would choose differently

I would reconsider an integration if many users asked for the same one and it could stay optional.

What changed as a result

The product became smaller in scope and easier to explain: capture tasks, let agents work in separate Git worktrees, and keep plans, changes, tests, and delivery decisions reviewable. A few people outside my own workflow have tried it since, which wouldn’t have made sense while it depended on my CRM.

Back to top