Nimble Plugin Adds Targeted Web Search to ChatGPT
The integration allows users to automate research tasks by applying entity retrieval and confidence scores to search results.
Updated on Oct. 29, 2026 in Artificial Intelligence
Live Poll
Do you trust AI-powered search tools to provide accurate and reliable information for your work?
Nimble has released a new plugin for ChatGPT and Codex that integrates its search system directly into these AI models. This tool, currently available to users, allows for targeted web research using entity retrieval to source specific data points.
Why it matters
The integration aims to reduce the overhead associated with repetitive web browsing and manual search tasks for AI users. By automating information retrieval, it seeks to improve the reliability of data accessible to language models.
Nimble’s Web Search Agents achieved a 74.0% accuracy score on the DRACO benchmark, which evaluates performance across 100 distinct cross-domain tasks. The system utilizes entity retrieval—a method that identifies and extracts specific objects or data from text—to execute targeted searches.
The players
Nimble
An AI firm specializing in autonomous web search systems and entity retrieval architectures.
OpenAI
The developer of ChatGPT and the Codex programming model which now support third-party search plugins.
The details
The tool functions by navigating websites and learning specific retrieval paths for recurring tasks, effectively refining its search process over time. Users can select proprietary or specific data sources for their workflows and apply confidence scoring to the results, allowing for a quantitative measure of accuracy. The system is designed to reduce the total number of browser calls required to satisfy a query.
Timeline
October 29, 2026: Nimble released the OpenAI plugin.
The Tech Race
Nimble’s integration represents a step toward standardizing how AI models verify information through web agents. It aligns with broader industry efforts to validate agent performance against specialized benchmarks like DRACO rather than relying solely on anecdotal model outputs.
Users can immediately access the plugin to integrate Nimble’s search system into their existing ChatGPT and Codex workflows. The tool allows developers and researchers to select specific data sources and set confidence thresholds for automated results.
The takeaway
The plugin moves the industry toward more verifiable AI research by allowing users to toggle source reliability via confidence scores. Watch for further updates to the DRACO benchmark as Nimble incorporates more complex cross-domain tasks into its search agents.
Further reading
Explore the broader ecosystem of Artificial Intelligence developments impacting modern search workflows.
Source note: This article includes information reported by IT Brief US.
Live Poll
Do you trust AI-powered search tools to provide accurate and reliable information for your work?
