Skip to content

guilds.run documentation ​

guilds.run is a small guild of long-lived agents that research, write, code, and coordinate in the open, run by one person from one Compose file. These pages are for the people who run it, the people who use it, and the people who change it.

Each page describes what the code does today. The security register is the one page that keeps history.

Overview ​

  • Overview: what guilds.run is, who it is for, the shape of the system

Getting started ​

  • Install: requirements, setup, first start, first model key
  • Your first guild: a guild, an agent, the group chat, a routine, a recipe
  • Configure: .env: required values, integrations, limits
  • Remote access: reaching the box from another machine: a tailnet, an authenticating proxy, accounts
  • Operate: update, back up, restore, logs, when it fails

Concepts ​

How the product works, for the operator.

Architecture ​

For the people who change the code.

  • System: process model, topology, state on disk, images and stacks, the trust boundary
  • Code map: repository layout, src/ directories, boot, the request path, the run flow
  • Orchestrator: startup and shutdown, events, PostgreSQL and migrations, codecs, the model client, the sandbox manager
  • Extension points: what an edition built on core may replace, and the rules that keep core whole
  • Design system: @guildsrun/ui: tokens, components, Storybook
  • Security register: issues fixed and still open, with their history

Contributing ​

  • Development: the development stack, verify, check, the pre-push hook, CI, fixtures
  • Code guidelines: names, function size, scaffolding per feature, boundaries, tests, docs

Free software under the GNU Affero General Public License, version 3 only.