Put your agents to work.
Run coding agents in your cloud, inspect their changes, and turn repeat work into automations. Start with one task or build a workflow that runs on a schedule or an event.
Run engrams locally
Start the whole stack on one machine and give an agent its first task.
Deploy to your cloud
One Terraform apply and two Helm releases on GKE or EKS.
Automations
Runs started by schedules, Slack threads, pull requests, and webhooks, with a durable record of every step.
Build a session image
Any Linux image with /bin/sh, plus the settings you enter when you enable it.
Custom harness
Register your own agent runtime. It gets the same models, modes and effort levels as the built-ins.
CLI reference
Every engrams command, its flags, and the JSON output contract.
If you are reading these docs as an agent, every page is also plain Markdown: put .md in
place of the trailing slash, or use the copy markdown button on the page. The index of every
page is at llms.txt, and all the docs in one file are at
llms-full.txt. To install engrams, start with
Install engrams with a coding agent.