T3 Code+20 Positive

GILD Review

GILD's call and the buying facts behind it.

From GILD research

Should we adopt T3 Code?

Should your engineers run their coding agents through T3 Code or through each agent's own app?

30-Second TakeT3 Code fits engineers who use more than one coding agent and want one app to run, compare and hand work between them.
Best fitEngineers and small teams who use agents from more than one vendor and want to run them side by side, hand work between them, and reach them from a phone.GILD Review
Consider another workflowIf everyone uses one vendor's agent, use that vendor's app; if you need SSO, audit logs or a support contract, ask T3 Tools first, since T3 Code is an MIT-licensed app that runs locally on each agent's own login with support through its Discord.GILD Review

Drawn from 73 sources and 1 operator insight

GILD analysis

T3 Code Review

T3 Code is a free, MIT-licensed app that runs the coding agents signed in on the user's machine, Codex, Claude Code, Cursor, Grok Build, OpenCode and Antigravity among them, from a desktop, browser or phone, with worktrees and cross-agent sub-agents. It is early: new threads start in Full access, it runs locally on each agent's own login, and Anthropic asks Agent SDK products to use API keys. Adopt it for engineers who mix vendors once permissions and Claude accounts are settled.

AI Coding Agents
Agentic Development
AI Costs

The decision

Best fit: Engineers and small teams who use agents from more than one vendor and want to run them side by side, hand work between them, and reach them from a phone.
Consider another workflow: If everyone uses one vendor's agent, use that vendor's app; if you need SSO, audit logs or a support contract, ask T3 Tools first, since T3 Code is an MIT-licensed app that runs locally on each agent's own login with support through its Discord.
Adoption condition: Change the default permission mode from Full access, keep telemetry as a conscious choice, and confirm how Claude and ChatGPT plans may be used through it.

Basis: T3 Code's README, docs, release history, privacy policy and terms, and Anthropic's and OpenAI's terms, read Nov 1, 2026. We ran no performance benchmark of T3 Code itself.

Buying & deployment factsWhat engineering leaders should verify before a broader rollout.

How you buy it

Free, MIT license

No paid plan; each agent bills through its own plan, login or key.

Source 1, T3 Tools, FAQ: FAQ

Details

Free under the MIT license; T3 Code's FAQ says there is no way to pay for it. Each agent bills through its own subscription or login, and API keys and OpenRouter work as an extra provider instance.

Cost drivers

Your agents' plans

T3 Code's FAQ says its Claude route drains 5-hour limits faster; T3 Connect is free for 3 machines.

Source 1, T3 Tools, FAQ: FAQ

Details

The agents' plans and keys are the cost. T3 Code's FAQ says its Claude route, through Anthropic's Agent SDK, drains 5-hour limits faster than Claude Code directly. T3 Connect is free for 3 machines per account.

Where it runs

Desktop, web, phone

macOS, Windows and Linux, a local web app, and iOS and Android remotes for a computer that stays on.

Source 2, GitHub, Install: Install Source 1, T3 Tools, FAQ: FAQ

Details

A desktop app for macOS, Windows and Linux, the t3 command with a local web app, and iOS and Android apps that control a computer that stays on. Agents run where T3 Code runs, a laptop, server or VPS.

What to verify

Full access default

New threads start in Full access, and telemetry is on by default.

Source 3, T3 Tools, Permission modes: Permission modes Source 4, T3 Tools, Telemetry: Telemetry

Details

New threads start in Full access, and T3 Code sends product usage data by default, tied to a hashed account or installation identifier and excluding prompts and file contents; T3CODE_TELEMETRY_ENABLED=false turns it off. Before rollout, confirm SSO, audit log and support terms with T3 Tools, and Claude plan use through the Agent SDK with Anthropic.