The open-source agent IDE
Clave
Every agent session is a terminal. Clave gives the terminals a home: tabs, groups and whole workspaces you launch in one click, whichever agent you run.
Clave began as a plain observation: every agent session is a terminal, and terminals deserve a better home than a window you lose.
So Clave is a desktop app that hosts terminals. Your agents, whichever you already use: Claude Code, Codex, Gemini, anything that runs in a shell. Sessions group by project, whole workspaces launch in one click, and a live page can sit right beside the sessions producing it.
We work inside it all day, on everything. It is completely open source, and deliberately unaware of anything else we build: our own tools drive it from the outside, and yours can do the same.
The Antasphere team
Bring your own agent
The agent layer stays yours. No wrapper around the model, no lock-in above it: if it runs in a shell, it runs in Clave, and switching agents never means switching homes.
Projects, not tabs
Sessions organize the way work does: grouped by project, side by side. A group remembers its sessions, its directories and its saved terminals, so clicking it raises the whole workbench, dev servers included, each session primed with the prompt and the context you declared for it.
An IDE that builds itself around you
Clave ships an MCP server, so your agents can drive the app they live in: open a sibling session, hand a task to a fresh agent, attach a live page to a group. Pin a local server to a group and clicking it shows the running page, a dashboard, a preview, a board, exactly where its sessions live. Let your agents generate those pages and the IDE starts building itself around your work: the interface stops being something you installed and becomes something your project grew.
Open source, all of it
The entire app is open source, not a free tier around a paid core. Fork it, extend it, wire your own tools into it: the MCP surface our products drive it through is the same one yours get.