Frequently asked questions

Yes. Override the agent per column, so a task plans with Claude, executes with Codex, and gets reviewed by Gemini on one board. Turn on handoff context for a column and the next agent picks up where the last one left off. Any agent can receive a handoff; whether the conversation carries over depends on the outgoing agent leaving a readable history behind, and the agent table says which do.

No. Kangentic seeds a default board on first launch; from there the workflow is yours to define. Rename, rewire, and add columns until the board matches the way your team ships, with only the To Do and Done endpoints fixed. Every column carries its own agent, model, effort, permissions, and auto-command, and it all commits with the repo in kangentic.json, with kangentic.local.json for personal overrides.

Yes. Your columns, automation, and per-column agent settings persist to kangentic.json at the project root, which commits with the repo, so a teammate who clones the project gets the workflow you shaped instead of a default board. Personal deviations go in kangentic.local.json, which Kangentic adds to .gitignore for you.

Kangentic is free, but the CLIs it runs bill your own provider accounts, so it tracks the spend for you. The Usage Stats dashboard reports tokens, cost, and burn rate for one project or the whole machine, broken down by agent, model, and reasoning effort. Each session's context bar shows its own cost and context use live, so an expensive run is visible while it is still running.

Yes. The Kangentic Mobile companion app pairs with your desktop over an end-to-end encrypted link so you can read agent transcripts, answer permission prompts, and review diffs from your phone. For current availability on iOS and Android and the steps to install, see how to get the app.