Tick
Deterministic task tracking for AI-assisted coding
Start new thread
Lee Overy

Tick - Deterministic task tracking for AI-assisted coding

by•
Tick is a lightweight task management CLI designed for AI coding agents. It prioritises determinism, simplicity, and zero-friction git integration. A fast Go CLI that tracks coding tasks in a machine-readable way so AI agents can reliably understand progress. Tasks are stored as a simple JSONL ledger and exposed / readable via a SQLite cache store. Outputs pretty in the terminal for humans, and TOON for AIs.

Add a comment

Replies

Best
Lee Overy
Maker
📌
Hi! I’m the developer behind tick. I built it while doing daily AI-assisted coding. I started with a TODO.md so the agent could resume work the next day. It mostly worked… until it didn’t. The AI would interpret the text differently each run, miss tasks, or reread everything and waste tokens. What I actually needed was: a simple file I could trust as the record of work, but tooling that behaved predictably. So tick works like this: The tasks live in a plain JSONL file — that’s the source of truth. But when you use the CLI, it loads that ledger into a local database cache and you interact with it like a proper task system. So you get: • a portable, inspectable task history • deterministic state for AI workflows • fast queries and updates • no heavy services or setup The database is ephemeral and rebuilds instantly from the file if missing. Nothing special to run and nothing required to commit. Typical use case: You stop a coding session → close your laptop. Tomorrow you run your AI again → it knows exactly what work is still incomplete and continues instead of starting over. tick is intentionally small and local. It’s not for teams or project management — it’s for a developer working alongside an AI in a repo. If you try it I’d really like to hear what agent/workflow you used and what felt awkward. I’m actively iterating on it.