komradGuides
These guides explain how komrad works and how to use it, written to be read start to finish. The API, CLI and config pages are generated from the code and say exactly what each call, command and key does. The guides say why, and how the pieces fit together.
Getting started
Install komrad, then build up from a workflow to agents that talk to each other.
- 01Installing komradGet the binary, give it a database, bind the panel key.
- 02WorkflowsThe templates agents are created from, and what each one decides for them.
- 03AgentsWhat an agent is, what the panel shows about it, and how agents work together.
- 04HooksYour own scripts, run by komrad at moments in an agent's life.
- 05SchedulesStart agents on a timer, and what happens when a run is late or busy.
- 06MailHow agents send each other messages, and the limits that keep them in check.
- 07MCPHow agents see and manage other agents through komrad's MCP server.
Operations
Connecting forges and running komrad on a machine of its own.
Architecture
How komrad works and why, for anyone curious or changing it.
- 11How it worksThe service, the panel, tmux, and the one API everything goes through.
- 12PhilosophyAn overview of the design decisions behind komrad, the terminal-based agent orchestrator.
- 13Events and the watchHow a committed row becomes a redraw, and how the loops keep the database honest.
- 14The remote engineHow we run komrad on a machine of its own, and what survives what.
- 15From proto to every surfaceOne RPC declaration becomes a handler, a CLI command, an MCP tool and a reference page.
- 16The shape of the codeThe layers, what may import what, the linters that hold the line, and where tests live.