bise docs
bise is a terminal app for multi-agent coding. you talk to one agent, main, and it runs a team of agents for you.
bise runs in your terminal, one window per repo. you talk to main, your team lead. main turns what you ask into tasks and starts one agent per task. the agents work in parallel, talk to each other, and land their work in your repo. what needs you waits in your inbox.
these docs cover how to install bise, how to give it a model, and how to use and extend it.
start here
- install bise and run it in a repo.
- give it a model: an API key from a provider.
- ask main for something. read main and the agents to see what happens next.
curl -fsSL https://bise.dev/install | sh
cd your-repo
bisewhat's in these docs
| page | what you find there |
|---|---|
| providers and keys | the providers bise knows, where a key goes, bise login |
| models and roles | which model does what: main, agents, small jobs, voice |
| gateways and local models | LiteLLM, a company proxy, Ollama, LM Studio, vLLM |
| main and the agents | how a task runs, the inbox, the commands you use |
| approvals | yolo and auto: what runs without asking |
| landing work | how agents commit, features, pull requests |
| voice | dictation and voice mode |
| computer use | agents that drive Chrome and your Mac apps |
| plugins and MCP | skills, local and remote MCP servers, logins |
| config.toml | every setting in one place |
| troubleshooting | bise doctor, logs, the usual fixes |
for agents
every page is also plain Markdown: add .md to its address, for example bise.dev/docs/providers.md. to have a coding agent install and set up bise for you, give it bise.dev/setup.md.
requirements
bise runs on macOS, on Apple silicon and Intel. you need git. linux is next.