Caffold is a self-hosted workspace for running and reviewing coding-agent work from any screen. Keep Codex or Claude Code working on a Mac you control, then follow the conversation, answer approvals, inspect commands and tests, and read the actual files and diff from a desktop, foldable, tablet, or phone.
The layout adapts to the screen; the workflow stays the same. A Task still contains one agent conversation and the repository context needed to judge its work. Start on one device, leave the turn running on the Mac, and return from another to decide what happens next by text or voice.
Caffold is not a hosted agent, remote terminal, or replacement harness. The agent CLIs, repositories, credentials, conversations, and execution remain on your Mac.
Follow a Task as it runs, then read the result and decide what comes next.
Open Working Tree to review the actual files and diff without leaving the Task.
These deterministic showcase images use a Codex Task. Claude Tasks use the same Conversation and review workspace.
When you create a Task, choosing a model also chooses the agent that provides it. Caffold currently supports:
- Codex, through its persistent app-server runtime; and
- Claude Code, through its CLI protocol and a Caffold runner that keeps the CLI process attached while the backend is replaced.
A Task remains bound to that agent for its lifetime. Its model, reasoning or effort choices, permission modes, tools, session behavior, and transcript come from the selected agent rather than from a Caffold reimplementation.
This is deliberate. A coding agent is the model together with the harness its authors built around it. Caffold gives Codex and Claude separate native drivers so it can preserve those harnesses instead of forcing both through a lowest common denominator. It normalizes only the product concepts the workspace must present consistently: conversations, turns, activity, approvals, and the operations a Task can actually perform.
See Agent runtimes for the design and the different state, transport, and recovery boundaries of the two integrations.
Only one machine does the actual work. Install Caffold Server.app on the Mac
that has the agent CLIs, Git, and your checkouts. The app stays in the menu bar,
keeps the Caffold backend available, and connects the browser interface to the
selected agent and the files on that Mac.
browser or installed PWA
(desktop, foldable, tablet, phone)
|
local URL or private
Tailscale HTTPS URL
|
Caffold Server on Mac
/ \
Codex app-server Claude runner -> claude CLI
\ /
Git checkouts and worktrees
On the host Mac, Open Caffold opens the local address. On another device,
Tailscale Serve can provide a private HTTPS address. Configure and copy that
address from Settings → Remote Access; the ready page also provides a QR
code for another permitted device.
Each browser or installed PWA is a window onto the same Mac, not another copy of the server. Closing it does not end a turn, but the Mac must stay awake, running Caffold, and reachable for remote use.
- Start a Task in the directory where the work belongs and choose a Codex or Claude model.
- Follow the conversation and answer the agent's approval requests while it works.
- Read the result, command and test output, changed files, and actual diff.
- Type or dictate the next instruction, steer an active turn, or return later and continue the same Task.
For longer turns, each browser can opt in to system notifications under Settings → Notifications, for a turn that ends and for a Task waiting on an approval.
Caffold also keeps the Task connected to its repository, optional managed worktree, Git history, and read-only GitHub Issue or Pull Request context.
Caffold supports Apple silicon Macs running macOS 14 or later. Install and sign in to at least one supported agent:
- the official standalone Codex CLI
0.147.0or newer; or - Claude Code
2.1.236or newer, available asclaudeon the app'sPATH.
Both may be installed, and Caffold will offer the models it can reach. The installation guide has the agent-specific setup and diagnostics.
Install Caffold with Homebrew:
brew install --cask panarch/tap/caffoldLaunch Caffold Server from Applications, then choose Open Caffold from its
menu-bar menu.
The Task composer supports host-local multilingual voice input. On first use,
Caffold asks before downloading the pinned Whisper large-v3-turbo model
(about 1.5 GiB). Audio is sent only to the same Caffold host, processed in
memory, and never stored or sent to an external speech service.
Voice is useful here for the same reason the browser interface is useful: much of the work is giving direction, reading what happened, and following up. A keyboard is welcome, but it should not be required for every step.
Caffold assumes one trusted user, one trusted Mac, and a private network. It does not provide authentication for a public deployment or multi-user authorization.
Agent support is built in rather than loaded as a runtime plugin. Caffold currently drives Codex and Claude Code; it does not provide an ACP driver or let an existing Task switch agents.
Its Git and GitHub views are deliberately review-oriented. Caffold does not provide a full editor or terminal, and it does not expose stage, commit, checkout, merge, rebase, reset, stash, publication, or review mutation controls. Those operations can still be requested through the Task's agent or performed with the developer tools you already use.
Managed-worktree preparation is explicit, and Caffold only cleans up worktrees that it created and recorded. The complete implemented scope and limitations are tracked in Current Product Status.
- Installation and operation
- Product vision
- Current product status
- Product workflows
- Agent runtime architecture
- Roadmap
- Contributing
- Testing
The complete documentation index groups product, architecture, development, review, and maintainer material by purpose. Caffold is available under the Apache License 2.0.