LiveShortly docs

LiveShortly streams a coding agent's session to a web page while it runs. People you share it with can watch, send the agent messages, and approve its tool runs. When the session ends, the page becomes a replay. New here? Start with Getting started.

1. Getting started

   Install the live CLI, sign in, stream your first agent session, and open it on your phone. You need macOS or Linux (Windows through WSL 2), a Google account, and a coding agent such as Claude Code, Codex, Gemini CLI or Antigravity installed and signed in on the same machine. It takes about five minutes.

2. Watching & commenting

   Watch a session live in the browser or the terminal, and send messages to the agent while it runs. You need read access to the session (see Sharing & visibility); sending messages needs comment access and a signed-in account.

3. Approvals

   Approve or deny an agent's tool runs from a tmux popup, the session page, your phone, or another terminal. Permission control is on by default for live claude, live codex, live agy and live gemini, and it fails closed: if nobody answers, or approvals can't be reached, the tool does not run. You need to own the session, or have comment access and use live join.

4. Sharing & visibility

   Decide who can open a session and who can send messages to it. You need to own the session. New sessions are private.

5. Publishing to the feed

   Put a session in the public feed so anyone can find it, watch it live or replay it, and take it out again. You need to own the session. A session can be published while it is live or after it has ended.

6. Forks & handoff

   Continue any session you can read as a new session that you own, with the same agent or a different one. The original session is not changed. You need read access to the source session and the live CLI signed in to the same server.

7. Starting sessions from the web

   Start agent sessions from the browser on a machine you control, using the live daemon. You need the live CLI installed and signed in on that machine, tmux installed, and the agent CLIs installed and signed in there too. Sessions run on your machine, not on the server.

8. Agents

   Which agents work with LiveShortly, what each one supports, and how to set each one up. You need the agent's own CLI installed and signed in; live wraps it and does not replace it.

9. PocketAI apps

   Use the PocketAI apps for iPhone and Mac to run agents on your own machine and approve their tool runs from your phone. PocketAI is made by LiveShortly; see pocket.liveshortly.com. You need a machine with live and your agent CLIs installed and signed in, and a network path from your phone to that machine.

10. Self-hosting

    Run your own LiveShortly server with Docker Compose, so sessions stay on infrastructure you control. You need a Linux or macOS machine with Docker Engine, the Docker Compose plugin and Git, a Google OAuth client for sign-in, and, for a public server, a domain name and a TLS certificate. Self-hosting is free; the server is open source.

11. CLI reference

    Every live command and flag. This page is generated from live --help; run live --help to see the same text for the version you have installed.

12. REST & SSE API

    The HTTP API behind the web app and the CLI. Use it to build your own client or automate sessions. You need a LiveShortly server and, for most calls, a signed-in user.

13. Troubleshooting

    Each message the CLI or the site shows, what causes it, and how to fix it. Start with the log: the CLI writes everything it can't show you to ~/.liveshortly/live.log, one line per problem, prefixed with the first 8 characters of the session id.

14. FAQ

    Short answers to common questions. Each links to the page with the details.

15. Glossary

    Terms used in these docs, in the CLI and on the site, in alphabetical order.