How to Add Context to Your Running ChatGPT Deep Research & GPT-5 Pro Query (Step-by-Step)

OpenAI has announced a new ChatGPT feature that allows users to interrupt long-running deep research or GPT-5 Pro queries and add new context without restarting or losing progress. If you've ever launched a ChatGPT deep research run or a heavy GPT-5 Pro query, only to remember an important constraint or context halfway through, you no longer have to cancel the task, tweak the prompt, and start over. The ChatGPT update allows you to interrupt a long-running query, inject new context, and the model will adjust your changes into its ongoing reasoning.

The majority of people online have liked this simple yet profoundly impactful update that gives users the ability to interrupt and update queries in real-time.

What's changed and why it matters

Long, multi-step queries are effective for combining information, resolving conflicting details, and creating organized outputs. However, they can easily become outdated when new data comes in, the focus changes, or you forget to request a CSV with the memo. The updated workflow turns these challenges into a beneficial feature.

Tap "Update" in the sidebar, add your new requirements, context, exclusions, and formatting preferences, and the model recalibrates without losing progress. It's faster, cheaper, and closer to how real research, product discovery, and analysis actually happen.

Here's a breakdown of what makes this update so significant:

• On-the-Fly Adjustments: Users can now interrupt a long-running query mid-stream to add new information or clarify a request. Instead of starting over, you can simply hit "update," add the new detail, and the AI will pivot its response accordingly.
• No More Lost Progress: The most significant advantage is the preservation of context. The AI retains all the information and nuances of the conversation, allowing you to integrate the new information seamlessly. This is particularly beneficial for deep research tasks or iterative processes where the context is built over a series of interactions.
• Increased Efficiency: This feature saves valuable time and reduces the friction often associated with complex AI queries by eliminating the need for constant restarts, streamlining users' workflow, and improving efficiency.
• A More Natural Interaction: This update makes interacting with AI feel less like a command-line interface and more like a conversation with a human assistant. The ability to interject, clarify, and refine our requests in real-time mirrors the natural flow of human dialogue, making the entire experience more intuitive and less robotic.

How to add context to your running ChatGPT deep research query:

Step 1: Using this new feature is straightforward and easy. Visit ChatGPT.com and start a deep research query on any topic.

Step 2: For the purpose of this article, I went with something simple and asked about "The recent growth and rise of agentic AI browsers."

Step 3: Once you have initiated the query, a sidebar will appear, showing you the research in process, which you can opt to stop or update.

Step 4: I clicked on update to add additional context to the query. I simply added, "Also, talk about how traditional browsers are still more useful and capable than the agentic ones for now." ChatGPT will understand the intent and add additional context to the search.

The entire process is similar for GPT-5 Pro.

In Conclusion:

The introduction of real-time query interruption and context updates turns AI from a batch process into a live workspace. The ability to add long-running queries and add new context to deep research and GPT-5 Pro tasks allows users to save time and better align with changing goals. The workflow is simple: click the Update option in the sidebar, add your clarifications, and let the model adjust. This provides users with clearer outputs and a research process that ultimately aligns with your thinking.