Both are free, MIT-licensed apps that run your own coding agents on your own plans. T3 Code is a control surface with desktop, web and phone apps, built to drive agents on any machine from anywhere. Friday stays on one computer and adds a shared task board with checked tasks, agents that message each other, and a hand-off when a tool runs out of usage.
What is Friday?
Friday is a free, open-source (MIT) harness for coding agents that work as a team. It runs on your own computer with no account: the desktop app for macOS, Windows and Linux, or one command with Node.js 22 and Git.
- Claude Code, Codex, Cursor and 12 more coding tools (15 by name, plus any ACP agent) run side by side on their own plans and sign-ins.
- One agent splits a job into a task board, the others pick up tasks, and every finished task is checked before it counts as done.
- Agents message each other as they work, and each can work in its own copy of the repository.
- When a tool runs out of usage mid-task, the task moves to the agent best placed to finish it.
- Review each run’s changes line by line, with git built in.
- No account; anonymous usage counts you can turn off; it listens on 127.0.0.1 only.
Install it with npx github:splashxmoon/friday or from GitHub; see the docs. Friday Team, the hosted version for several people in one session, is available to companies: talk to us.
How do Friday and T3 Code compare?
| Friday | T3 Code (from its own repo and site) | |
|---|---|---|
| Price | Free, MIT license | Free and open source (MIT); you bring your own subscriptions |
| Platforms | Desktop apps for macOS, Windows and Linux | A desktop app and command line for Mac, Windows and Linux, a web app, and iOS and Android apps |
| Coding agents | Claude Code, Codex, Cursor and 12 more, plus any ACP agent | Claude Code, Codex, Cursor, OpenCode, Antigravity and Grok Build, and more listed in its docs |
| Where it runs | On your own computer; it opens only there (127.0.0.1) | On your computer or a server you run, with remote access from a phone or browser |
| Keeping agents apart | Each agent can work in its own git worktree, with a warning when lanes in separate copies change the same file | Git worktrees, with automatic cleanup in its settings |
| Agents working together | A shared task board with checked tasks, and agent-to-agent messages | Subagent activity shown in the thread, and an MCP server so other agents can start and message threads; a shared task board and agent-to-agent messages are not described in what we checked |
| When a tool hits its usage limit | The task moves to another agent with a record of what it did and changed | The thread shows Limited; you resume at the reset or switch provider yourself, and recent conversation context goes with you |
| Review and git | Line-by-line review with notes sent back to the agent | A diff pane for each turn, pull request review and merge, and review guides (beta) |
| Automation | Crew teammates with roles and memory, and workflows with a check step and your OK before a step | Scheduled tasks and webhook automations |
| Account and usage data | No account; anonymous usage counts you can turn off | Sends basic product usage events, off with T3CODE_TELEMETRY_ENABLED=false; your code and prompts go to your provider, not to T3, per its FAQ |
Last checked: 2 November 2026
Based on T3 Code’s own repository and site, checked 2 November 2026. Check t3.codes and github.com/pingdotgg/t3code for its current features.
When is T3 Code the better pick?
If you want to run agents on a remote machine or a VPS and drive them from your phone or a browser, T3 Code is built for that, with iOS and Android apps and its own remote connection. It also has scheduled tasks and webhook automations, pull request review and merge inside the app, a guest link to let a teammate follow one thread, and an MCP server so other agents can start threads. It is a project with a nightly build every day, and a large community around it. If one person driving many agents from anywhere is the job, start there.
When is Friday the better pick?
If you want one agent to split a goal into a task board that other agents pick up, a second agent to check each finished task, or Claude Code and Codex to ask each other questions instead of you relaying them, Friday fits. It also fits if you want everything on one computer, opening nowhere else, with Work for tasks that are not code and Crew for teammates that keep roles and memory.
How does usage-limit handling differ?
T3 Code’s docs describe it this way: when a provider stops on a usage limit, the thread shows Limited, and you can schedule the thread to resume at the reset, or switch to another provider yourself. When you switch, it passes a budgeted selection of the recent conversation, not the first agent’s tool state or attachments.
In Friday, the task moves to another agent when a tool runs out, with a record of what the first agent did and changed, built from its session log. They are different trade-offs: in T3 Code you decide where the work goes; in Friday the task moves on. We measured one pair of agents in the hand-off benchmark, which was run on the hosted app and says so.
Can you use both?
Both are free and both run the coding tools you already have on your own plans, so you can try each on the same repository. For how parallel agents stay apart on one repository, see how to run coding agents in parallel.