Bica
Bica runs coding agents on the machine that holds your code and gives you one web page to work with them. You give an agent a task, watch it work, steer it while it runs, and review what it changed.
Each agent works in a session, and you can run many sessions at once. A feature can get its own git worktree, so agents working in parallel don’t touch the same files. You read their changes in full-size diffs and commit them from the Git tab. Closing the browser stops nothing: the agents keep working, and a run that was cut off by a restart picks up where it stopped.
Bica is one program, bica. Your code, your sessions, and your sign-ins stay on your machine, and only the calls to the model leave it. You pay for the model with a Claude or ChatGPT plan you already have, or with an API key.
How it differs from Claude Code or Cursor
Section titled “How it differs from Claude Code or Cursor”Claude Code and Codex run in a terminal, and Cursor is an editor with an agent inside. Bica is a web page for running several agents side by side and reviewing their work. You still open your editor for heavy editing.
Bica runs its own agent and doesn’t drive Claude Code underneath. It reads the AGENTS.md or CLAUDE.md files and the skills you already have. Since the agents run on your machine and not in the page, you can start a task on the laptop and follow it from your phone.
Where to go next
Section titled “Where to go next”Read Start here from the top: install Bica, send a first prompt, and get a picture of how it works. Then pick the pages you need. The FAQ answers the usual first questions in a line or two each.