Termdeck

The agents are running. Go live your life.

Web control plane for the coding-agent CLIs already on your machines. Claude Code, OpenAI Codex CLI and Grok: one fleet board, every device, tool approvals on your phone.

termdeck.io · Landing page · Install · How it works · FAQ

The minute after

The agents already run fine. That was never the hard part.

The hard part is the minute after. A run stops on a tool approval and sits there until you happen to wander back to the terminal. Overnight runs die at 2am on a one-tap decision. Three sessions across two machines become three terminals you have to remember to check. A laptop lid closes and the SSH session goes with it.

Termdeck is the board that watches all of it, and pushes you the one moment that needs a human.

Disk is the source of truth

This is the design, and it is the one claim a wrapper cannot copy.

Termdeck does not run agents in its own sandbox and it does not keep its own copy of your conversations. It renders the CLI's own session files, on your machine: ~/.claude/projects/<slug>/<id>.jsonl, $CODEX_HOME/sessions, ~/.grok.

Three things fall out of that, for free:

• Every session shows up. Not just the ones you started through Termdeck. Add a project folder and its whole terminal history is already there, months of it, searchable.
• Round-trip parity is 100%. A session you unblocked from your phone on the train still opens with claude --resume <id> at your desk. Native storage, real cwd, no export step.
• A terminal-attached session streams live, in view-only, with a Take over button that closes the idle terminal client and hands you the composer.

Three engines, one UI

	Claude Code	OpenAI Codex CLI	Grok
Render every past session	●	●	●
Drive turns from the browser	●	●	●
Inline tool approvals	●	●	●
Steer a running turn mid-flight	●	●	○
Model + reasoning effort per turn	●	●	●
Native plan mode	●	●	○
Resumable in its own CLI afterwards	●	●	●

The engine is picked per chat. They share one sidebar, one transcript renderer and one permission card, so there is nothing new to learn when you switch.

What you get

One fleet board	Every machine you connect: laptop, desktop, VM, the box under the desk. Sessions sorted by who needs you, not by which terminal they live in.
Approve from your phone	Web Push fires with no browser open. Allow or Deny on the notification itself, or tap through to the session.
One approvals inbox	Every request waiting on you, across chats and machines, in one list you can answer in bulk.
Answer the agent inline	AskUserQuestion renders as real options. No switching to a terminal to type 2.
Take over a terminal session	The view-only lock, released. Close the idle CLI client and drive from the web.
Switch model and effort mid-run	Per turn. Drop to a cheap model for the mechanical part, jump to the big one for the hard part.
Docked diff review	Changed files with +/− counts, next to the transcript, and a pull request opened from the same dock.
Terminals canvas	Every chat as a live pane on one grid, across machines.
Background shells	A dev server a chat started keeps running after the reply, and the chat picks up when a shell finishes.
Fleet-wide search	Every session on every machine, by prompt, project or host.
Cost and pace meters	API-equivalent cost per project, token totals, cache-read rate, and rolling and weekly limit pacing before you hit the wall.
Switch accounts	Saved engine logins per machine, each with its own usage meter.
Install to your home screen	PWA with a push subscription per device. The device you are looking at keeps the others quiet.

Install

Two minutes, and no inbound port on your machine.

1. Sign up at termdeck.io/signup with GitHub, or email and password, and start the 14-day trial.

2. Add a machine (Settings, then Machines and accounts). You get a one-liner carrying that machine's token:

curl -fsSL https://termdeck.io/install.sh | TERMDECK_AGENT_TOKEN=agt_… sh

$env:TERMDECK_AGENT_TOKEN="agt_…"; iwr https://termdeck.io/install.ps1 | iex

Or with npm, the same install on any platform (termdeck-agent):

npx termdeck-agent install --token agt_…

npx termdeck-agent status, logs, repair and uninstall cover the rest of the agent's life.

3. Add a project folder. Its sessions are already there.

macOS, Linux and Windows. The agent installs as a user service (launchd / systemd --user / schtasks, with a @reboot cron fallback) and keeps itself updated. Full unit templates in deploy/AGENT-SETUP.md.

How it works

The agent dials out over a reverse WebSocket. There is no inbound port, no firewall rule and no tunnel on your side. Turns run on your machine: the agent spawns the CLI, reads its stream and keeps a numbered log, so a dropped link or a restart of the master is a viewer going away, not a lost run. The agent and the browser time termdeck.io and edge.termdeck.io and connect through whichever answers faster. If the one in use fails, they move to the other at once.

your machine                          termdeck.io
┌────────────────────────┐            ┌──────────────────────────┐
│ ~/.claude  ~/.codex    │            │  master                  │
│      ▲                 │            │   ├─ session list cache  │
│      │ watch + read    │            │   ├─ browser WebSocket   │
│ ┌────┴─────┐           │            │   └─ Web Push            │
│ │  agent   │ ──── reverse WS ────▶  │                          │
│ │  index,  │           │            │                          │
│ │  runs    │           │            │                          │
│ └────┬─────┘   (dial-out, TLS)      └───────────┬──────────────┘
│      │ spawn                                    │
│  claude / codex / grok                      your browser
└────────────────────────┘                    (phone, laptop, tablet)

The agent is in this repo. Read it before you run it. That is why it is here. It is a fixed list of typed capabilities, not a remote shell: reads confined to the transcript roots and the open chat's project folder, a few narrow named writes, processes it started itself. It redacts its own token out of its logs, never logs anything it serves, and stages every self-update behind a compile check with a .rollback/ snapshot and two watchdogs.

Start at agent/agent.js and agent/capabilities.js. The second one is the whole list of what a machine will answer.

Built in the open, and written down

53 load-bearing invariants, each one carrying the failure that produced it: docs/INVARIANTS.md.

A preview is a PAINT, read from the TAIL. Seeds and ?tail=1 parse only the file's last bytes; startLine keeps totalLines exact, because it is the cursor every ?since splice rides.

A delta may only be spliced onto the cursor it was fetched at. A live push landing inside a ?since fetch is what duplicated the transcript, on screen and in the cache, whenever a second chat was live.

3,056 commits · 1,013 production deploys · 112 days 738 test files · 53 invariants · 0 build steps

Compared to

	Termdeck	Happy	Claude Remote Control
Claude Code	●	●	●
OpenAI Codex CLI	●	●	○
Grok	●	○	○
Sees sessions it didn't start	●	○	○
Push approval to phone	●	●	●
Take over a terminal session	●	○	○
No extra subscription	○	●	●

Anthropic's Remote Control is good and it comes with your Claude plan, if you run one engine and only the sessions you switch it on for. Happy is free, open source and end-to-end encrypted, for sessions you start through its wrapper. Termdeck is for every session, on every machine, in all three engines. Longer write-ups: termdeck.io/vs.

FAQ

Can I approve Claude Code permissions from my phone?

Yes. That is the feature the product exists for. When a turn blocks on a tool approval, Web Push fires with no browser open. Tap Allow or Deny on the notification, or open the session for the full card. The run continues from your pocket.

Is there a self-hosted version?

No, and deliberately. There is one way to run Termdeck: connect a machine to your account and drive it through the cloud. Your code and your transcripts never leave your machine (the agent reads and renders them in place), but the control plane is hosted.

Does a session I drive from the browser still work in the terminal?

Yes. claude --resume <id> and codex resume <id>. Native storage, real cwd, and Codex resumes the thread before every web turn so terminal appends are picked up. Round-trip parity is a hard rule, and it is tested.

Does it work with Codex CLI?

Yes. Full turn driving over one shared codex app-server per machine, including native plan mode, with Claude Code and Grok in the same sidebar.

What can the agent actually touch?

A fixed list of typed capabilities, not a shell. Reads are confined to the transcript roots and the open chat's project folder (read only). Paths are never sent from the server: the master names a session id and the agent resolves it against its own disk. The writes are narrow and named: a file you attach (into ~/.termdeck/uploads, never your project), the project instruction file, checkpoint restore, and the engine account switching you ask for. Processes it may signal are ones it started, plus a terminal client you asked to take over. The code is in agent/capabilities.js; termdeck.io/docs/security is the long version.

The coding agent it drives is another matter: in full access mode it can do anything you can. That is what the permission rules are for.

What does it cost?

Starter is $9.99 a month for 1 machine, Pro is $19.99 for 10 machines with unlimited projects and saved logins, your whole usage history and a bigger terminal grid. Every account starts with a 14-day trial of either; a card is needed and the first charge is on day 15. Details: termdeck.io/docs/plans. You bring your own Claude / Codex / Grok subscription; Termdeck never resells inference.

What's in this repo

This repo is the agent and the documentation: the parts that run on your machine, so you can read them before you trust them. The hosted control plane (session index, browser WebSocket, push, billing) is not open source.

agent/                 the thin agent: root-confined capabilities, self-update, log redaction
install.sh / .ps1      the one-liners the dashboard hands you, as termdeck.io serves them
deploy/                systemd / launchd / schtasks units, heal + uninstall scripts
docs/ARCHITECTURE.md   how the pieces fit
docs/INVARIANTS.md     53 load-bearing rules and the failure behind each one

agent/ is the installed agent

agent/ holds exactly the files termdeck.io serves to an installing machine, laid out flat the way they land in ~/.termdeck/agent/. Some of them (the transcript parsers, the session index, the disk primitives) are shared with the master, which runs the same modules: one copy of the rules, because two copies of a transcript parser drift and a drifted parser renders a chat differently depending on which side read it.

Every release is signed offline (agent/update-signature.js), and an installed agent takes an update only when the signature and every file hash check out.

Your agents are already running.

Start at termdeck.io →

If this is the thing you keep wishing existed, a ⭐ helps other people find it.

_{Built by Bhavik Patel · termdeck.io · @termdeck}