# OpenAI — Greptile takes bug detection beyond code review with GPT-6 Astra

- Company: OpenAI (openai.com)
- Announced: 2026-10-26T22:00:00+00:00
- Category: not stated
- Coverage: not counted
- Announcement: no
- Group: routine
- Source: https://openai.com/index/greptile-code-review-codex-astra
- Record: https://forck.live/items/28023-greptile-takes-bug-detection-beyond-code-review-with-gpt-6-astra
- Subject: GPT / ChatGPT / API
- Models affected: GPT-6 Astra, Daybreak, Codex

Greptile is using OpenAI's GPT-6 Astra to test applications by running them, not just reading code, according to an OpenAI customer story. Co-founder and CEO Daksh Gupta says the model's computer use lets Greptile start a development server for a pull request, resize the browser to an iPhone-sized viewport and simulate a user moving through the app, so its visual understanding can catch problems such as a button that disappears on mobile. Gupta argues a QA firm testing every change is "just impractically slow" given how fast code now ships, and that GPT-6 Astra makes this kind of testing automatable at scale. Greptile also uses Daybreak to strengthen customer defenses, which Gupta says gives customers more sophisticated defense mechanisms than adversaries have offense mechanisms. Its Knowledge Base indexes a codebase and is exposed to Codex through an MCP server, and Gupta says thousands of customers have connected Codex to it; the team tracks how often AI-generated code is reverted and says fewer reverts happen when agents get context. Greptile also uses Codex to build its own software, and Gupta says he is most excited about Ultrafast as a way to move as fast as the team can come up with good ideas.

## Evidence

Verbatim from https://openai.com/index/greptile-code-review-codex-astra:

> How Greptile uses GPT-6 Astra to test user workflows, Daybreak to strengthen security, and Codex to reduce rework.

---

Record: https://forck.live/items/28023-greptile-takes-bug-detection-beyond-code-review-with-gpt-6-astra
Catalogue: https://forck.live/llms.txt
Current issue: https://forck.live/feed.md