AI NewsDev toolsAnnouncement

pingdotgg/t3code reaches trending rank 9 with mobile control for coding agents

T3 Code is an open-source dashboard for coding agents like Claude Code and Codex. The dashboard runs on desktop, web, iOS and Android.

AI News

Editorial2 min read

LinkedInX
T3 Code GitHub repository

Image: GitHub

Why it mattersA long agent run that needs an approval mid-way often stops dead if the person is not at the desk. A phone client fixes that.

T3 Code is a dashboard for running and controlling AI coding agents. It ranked ninth on Trendshift's daily list on 20 November 2026. The repository under pingdotgg has 67,861 stars. It was created on 8 February 2026. The project released v0.1.11 on 20 November with 643 changes from 138 contributors, by the release notes.

What it does

T3 Code is a single interface for running coding agents. It connects to Claude Code, Codex, Cursor and more. The user can start a thread on a desktop. The user can then approve the next step from a phone. The threads stay in sync across every client. The licence is MIT.

The project calls itself "the open-source control plane for coding agents". The code runs as a server on the user's own machine. The server talks to agents installed locally, authenticated as the user. Each thread runs in its own git worktree. The clients connect to the server over a local network, Tailscale, or T3 Connect, by the README's own account. There is a hosted web app at app.t3.codes for users who want to skip the server.

The iOS and Android apps are on Apple's App Store and Google Play. The desktop app installs through Homebrew, winget, yay or Nix.

What v0.1.11 ships

The release notes list "secret cards with show/hide toggles, shared Antigravity skills, thread failure reasons" alongside the 557 other fixes. The repository has moved fast all year. It has 1,300 open issues and 2,400 pull requests merged.

T3 Code does not resell tokens, the README writes. A user pays only for the model subscription or API key behind it, one for each agent.

The gap it fills

A long coding-agent run often sits waiting on a human for two minutes to approve the next step. A person who stepped away loses that whole window. The dashboard turns those two-minute gaps into something a phone can answer. The approval, the diff review, and the commit all happen in a chat-like thread. The user can also switch between different AI models mid-thread, per the README.

What a reader could do with it

A team already running Claude Code or Codex can install T3 Code. Point every agent at it. The team keeps the same subscriptions. The gain is the mobile and web client, and the shared thread history. The reviewer is still the team, on whatever device is nearest.

Source

SourceGitHub

This item was written by an AI system from the linked source. Reveneau is responsible for what it publishes.

Share
LinkedInX
Start a project