Science & Tech · AI · 5 hrs ago
A writer had Claude Code, Codex and Antigravity build one study app
An XDA writer had three AI coding tools collaborate on a study app built from scratch.
The tools were Anthropic's Claude Code, OpenAI's Codex and Google's Antigravity.
Each agent handled a different part of the project.
The writer had previously compared the tools and found that their strengths and weaknesses differ across tasks.
Codex was technically capable but needed extra guidance for the app's look, while Gemini surprised the writer with its design and product choices.
Using several agents can help cover gaps in one tool and let the writer switch if they reach a usage limit.
Tech journalist Mahnoor Faisal had Claude Code, Codex and Google’s Antigravity collaborate on a study app.
She assigned each AI coding agent a different part of the project.
Faisal said she wanted to use the tools’ different strengths rather than compare them as competitors.
She said no single AI model is best at every task.
- Who
- Tech journalist Mahnoor Faisal used Anthropic’s Claude Code, OpenAI’s Codex and Google’s Antigravity.
- What
- The three AI coding agents collaborated to build a study app from scratch, with each handling a different part.
- When
- Not stated
- Where
- Not stated
- Why
- Faisal wanted to combine the tools’ different strengths instead of treating them as competitors.
This story does not have two clearly opposing sides.
Faisal said she had been comparing AI coding tools and using them for different tasks.
Faisal began considering what might happen if the tools worked together instead.
Faisal had Claude Code, Codex and Antigravity collaborate on a study app from scratch.
- Writer
- Mahnoor Faisal
- Outlet
- XDA
- AI coding tools
- Claude Code, Codex and Antigravity
- Project
- A study app built from scratch
- Work split
- Each agent handled a different part of the project
1 source
What people are saying
I made all three collaborate to build a study app from scratch, with each agent handling a different part of the project.
there's no single LLM that's the best at everything.