OpenAI is taking a significant step forward with its AI-powered “deep research” feature by enabling codebase analysis on GitHub.
On Thursday, OpenAI revealed what they’re calling the first “connector” for ChatGPT deep research. This tool scours the web and other resources to create detailed research reports on various topics. The exciting news? ChatGPT deep research can now connect to GitHub (currently in beta), allowing developers to ask specific questions about codebases and engineering documents.
The connector will be available for ChatGPT Plus, Pro, and Team users in the coming days, with support for Enterprise and Edu users on the horizon, according to a spokesperson from OpenAI.
OpenAI’s ChatGPT Deep Research feature can now connect to GitHub
Image Credits: OpenAI
This GitHub connector arrives as AI companies strive to enhance the usefulness of their chatbots by integrating them with external platforms and services. For instance, Anthropic recently unveiled Integrations, which allows apps direct access to its AI chatbot, Claude.
Although OpenAI offered a plug-in capability for ChatGPT in the past, it shifted focus towards custom chatbots known as GPTs.
“I often hear from users that they find ChatGPT’s deep research agent incredibly valuable and want it to link to their internal resources, in addition to the web,” wrote Nate Gonzalez, OpenAI’s Head of Business Products, in a LinkedIn blog post. “[That’s why] today we’re introducing our first connector.”
Not only does the new GitHub connector answer questions about codebases, but it also empowers ChatGPT users to break down product specifications into actionable technical tasks and dependencies. Users can summarize code structure and patterns, and grasp how to implement new APIs with real code examples.
Of course, there is always a possibility of ChatGPT deep research generating inaccuracies—after all, no AI model is completely free from the occasional mistake. However, OpenAI is positioning this new feature as a time-saving tool rather than a substitute for human expertise.
A representative from OpenAI stated that ChatGPT will adhere to organizational settings, ensuring that users can only see GitHub content they are authorized to access and only codebases explicitly shared with ChatGPT.
OpenAI has been investing in its tools aimed at assistive coding. Recently, they launched an open-source command-line tool called Codex CLI and have upgraded the ChatGPT desktop app to read code from various developer-focused applications. The company views programming as a key area for its models. As a testament to this focus, reports indicate that OpenAI has reached an agreement to acquire the AI-powered coding assistant Windsurf for $3 billion.
In other news from OpenAI on Thursday, the company unveiled new fine-tuning options for developers looking to customize its latest models for specific applications. Developers can now fine-tune OpenAI’s o4-mini “reasoning” model using a technique called reinforcement fine-tuning, which leverages task-specific grading to enhance the model’s performance. Fine-tuning is also making its way to the company’s GPT-4.1 nano model.
Only verified organizations have the ability to fine-tune the o4-mini model, according to OpenAI. Meanwhile, fine-tuning for GPT-4.1 nano is open to all paying developers.
OpenAI started implementing verification requirements for certain models and developer features back in April. This process necessitates that organizations submit identification and other related documents, which the company claims is essential for preventing misuse.
