Understand this story
On October 26, 2026, Nace AI launched Drex for Search, applying its Drex decision model to document retrieval. According to Testing Catalog, the tool lets agents locate content layer by layer—by folder, file, and page—within cloud files, returning answers, citations, and page numbers; it can connect to AWS S3, Azure, and DigitalOcean, read scanned PDFs, spreadsheets, emails, and audio, and can be invoked as a tool call by coding agents such as Claude Code, Codex, and Hermes Agent. Nace claims a 62% reduction in steps per question and a 56% reduction in input tokens; across 529 queries on six professional work benchmarks, agents using the tool achieved 67.1% accuracy, up from 63.1% previously. MarkTechPost reported the next day that coding agents such as Claude Code, Codex, Cursor, and OpenCode can call it via CLI, SDK, MCP, or REST; paired with GPT-5 on BrowseComp-Plus, it achieved 92.05% accuracy with 12.61 search calls per question, 16.6 times fewer than the top-ranked entry; on VisDocAgentBench, with a Qwen 27B planner, it achieved 77.5% R@1, 10 percentage points higher than Claude Opus 5.
AI-generated from public reports
Latest update · Discussion
Testing Catalog joined this discussionDevelopments
1 confirmed developments
- Source:MarkTechPost
Reports and discussion timeline
Browse reports and discussion by time. Their evidence levels remain distinct.
Oct 28
Testing CatalogRelated discussion
Testing Catalog joined this discussion
This post contributes to heat; it is not a formal report or confirmed fact.
Oct 27
MarkTechPost
Nace AI Releases Drex for Search Retrieval Layer, Reaches 92.05% on BrowseComp-Plus
Nace AI has released Drex for Search, applying its Drex decision model to document retrieval. Coding agents such as Claude Code, Codex, Cursor, and OpenCode can invoke it via CLI, SDK, MCP, or REST. Paired with GPT-5 on BrowseComp-Plus, it achieves 92.05% accuracy with 12.61 search calls per question, 16.6 times fewer than the top-ranked entry. On VisDocAgentBench, using a Qwen 27B planner, it achieves 77.5% R@1, 10 percentage points higher than Claude Opus 5.
Oct 26
Testing Catalog
Nace AI launches Drex for Search retrieval tool
Nace AI has launched Drex for Search, enabling agents to locate content layer by layer from cloud files by folder, file, and page, returning answers, citations, and page numbers. The tool can connect to AWS S3, Azure, and DigitalOcean, read scanned PDFs, spreadsheets, emails, and audio, and can be invoked as a tool by coding agents such as Claude Code, Codex, and Hermes Agent. Nace claims a 62% reduction in steps per question and a 56% reduction in input tokens; across 529 queries on six professional work benchmarks, including Databricks OfficeQA, agents using the tool achieved an accuracy of 67.1%, up from 63.1% previously.
Heat trend for this story
- Current
- 11
- Period peak
- 19 Oct 27 01:00
- 24-hour change
- -12%
Move the pointer or select the chart to inspect hourly heat. Use the left and right arrow keys with a keyboard.