Features
Or the same agent, as a conversation.
Open it Integrated for permission cards, plans and diffs. The native CLI still does the work.
It starts helpers.
Agents start helpers, message them and wait for their reports. Each one nests under the agent that started it.
Your machines, working together.
Pair Patterm on your dev box and your desktop. Their agents can message each other and wait for replies, and you see every agent on both machines in one window.
It can draw you a chart.
Integrated agents can answer with charts, tables, diagrams and choices. They send the numbers, and Patterm draws them to match the app.
When it needs you, you’ll know.
Permission prompts, questions and finished work land in one list, and notifications can follow you to your phone.
Or ask what’s waiting.
A Codex-powered assistant keeps track of every agent in every project. Talk to it by text or by voice.
You review before anything lands.
Give a helper its own worktree, then read its changes file by file. Patterm never commits for you.
Plans that outlive the chat.
Todos with blockers and owners, and Markdown scratchpads that agents read and write too.
Every old conversation, in one place.
Your Claude Code, Codex and OpenCode history in one list, including sessions started outside Patterm. Search all of it from the top bar.
Know what’s left before you start.
Usage shows what’s left of your Claude Code and Codex 5-hour and weekly limits, when each resets, and the tokens you’ve used.
Work that carries on while you’re away.
Routines start an agent on a schedule with your instructions, and notify you when it reports.