OpenAI Adds GitHub Connector to ChatGPT’s Deep-Research Tool for Code Analysis

OpenAI has enhanced its AI-powered deep-research feature by introducing a GitHub connector, enabling ChatGPT to analyse codebases on GitHub.

Announced on Thursday, this is the first connector for ChatGPT’s deep-research tool, which compiles comprehensive research reports by searching the web and other sources. The integration is currently in beta testing, with ChatGPT Plus, Pro and Team subscribers gaining access in the days following the announcement. Enterprise and Education plan users will have access soon, according to OpenAI.

Users can grant ChatGPT access to specific code repositories and query their contents, receiving detailed responses with cited reports. The connector allows tasks such as breaking down product specs into technical tasks, summarising code structures and understanding API implementations using real code examples, all while respecting organisational permissions to ensure users only access authorised content. However, some early feedback on X highlights issues, with reports of inaccuracies or “hallucinated” responses, such as generating reports for unrelated issues.

This move aligns with OpenAI’s broader efforts to bolster coding capabilities, including the recent launch of the Codex CLI and reported negotiations to acquire Windsurf, an AI-powered coding assistant, for $3 bn.

The GitHub connector for ChatGPT deep research has been released as AI firms aim to enhance their AI-powered chatbots by integrating them with external platforms and services. For example, Anthropic has launched Integrations, which connects apps to its AI chatbot Claude.

Previously, OpenAI offered a plug-in capability for ChatGPT; however, it was abandoned in favour of custom chatbots known as GPTs.

Deep Research was launched in February as an agent capable of independently completing multi-step research projects. It collects, analyses and synthesises data from internet sources to generate detailed reports, saving significant time compared to traditional manual research. Following its release, a lightweight version was made accessible to free users, and those who exceeded their usage limits on subscription plans contributed to enhancing the accessibility of this feature. To process large amounts of web content, the programme employs complex reasoning models.

Deep Research

Deep Research is OpenAI’s advanced agent, designed to operate independently on behalf of the user. When given a prompt, ChatGPT searches, analyses and synthesises information from hundreds of online sources to produce a research-level report.

Powered by a specialised OpenAI o3 model optimised for web browsing and data analysis, it interprets vast amounts of text, images and PDFs, dynamically adapting to new information.

It is tailored for professionals in fields such as finance, science, policy and engineering who require precise, reliable research. Additionally, it benefits consumers seeking in-depth recommendations for products requiring extensive research, such as vehicles, appliances and furnishings.

Every output includes clear citations and a reasoning summary for easy verification. It excels at uncovering specialised, non-intuitive information that typically requires browsing multiple sources. By managing complex, time-consuming web research in a single query, Deep Research saves significant time and effort.