Run several coding agents from one interface

A developer uses T3 Code to manage multiple coding agents on personal machines, a VPS, and a separate work setup for code changes and writeups.

Done with Claude CodeCodexGitHub Copilot· Intermediate· Nov 2026

What they did
They use T3 Code as a shared control layer for Claude Code, Codex, and pi across their own computers and a VPS, while a different T3 Code setup at work connects to GitHub Copilot through ACP. Their usual approach is to split work into very small tasks, then raise the reasoning level or switch to a larger model when the job needs planning, design, or architecture.
How it went
That workflow has handled short coding tasks reliably, and DeepSeek 4.1 Flash has also done the job for them. The part that still needs the most human cleanup is prose, so they often rewrite reports, commit notes, and PR writeups themselves.
Tips
Break changes into narrow steps so the agent can stay on track. Save heavier models and higher reasoning settings for design-heavy or cross-component work, and plan to edit any generated writing.

Try it yourself with Claude Code

In [project folder], use [agent/model] to implement [task]. Start with a brief plan, then make the smallest safe code changes. If the work needs an architectural choice or affects several subsystems, stop and ask me before continuing. When the tests pass, prepare a short commit message and a pull request summary for my review. Do not commit, push, submit, delete, or send anything until I approve the diff and the text.

Fill in the [brackets]. Agents can make mistakes, so review anything before it is sent, paid or deleted.

Read the original on news.ycombinator.com ↗

Get 3 agent uses like this every morning

With the day's top AI news. Free, and one click to unsubscribe.