Nace AI's Drex for Search: hosted file retrieval for AI agents
Article timeline
The articles in this story. Times are ET.
🚨 AI News | TestingCatalogXDrex for Search tops Snowflake's MADQA leaderboard with 94.7% accuracyAINace AI's Drex for Search has reached the top spot on Snowflake's MADQA leaderboard, answering 94.7% of 500 document-based questions correctly. The tool lets AI agents find exact answers in company files and show the source page. It can be added to Claude Code, Codex, Cursor, and OpenCode as an MCP tool that agents call when needed.
- MarkTechPostNewsNace AI releases Drex for Search, a document search layer for AI agents
AINace AI has released Drex for Search, a hosted retrieval layer that coding agents such as Claude Code, Codex, and Cursor can call through CLI, SDKs, MCP, or REST. Nace says it scores 92.05% on BrowseComp-Plus with GPT-5 using 12.6 search calls per question, and 77.5% R@1 on VisDocAgentBench, 10 points above Claude Opus 5's 67.5%. Nace reports that in tests on OfficeQA Pro, agents using Drex for Search answered 6% more questions correctly with 62% fewer steps and 2.3x speed.
elvisXNace launches Drex for Search, reporting 62% fewer agent steps and 56% fewer tokensAINace has introduced Drex for Search, a decision-powered retrieval tool for agentic search through its File Search API. Nace reports 62% fewer steps to find answers, 56% fewer tokens, and a 4% performance gain across Claude Code, Codex, OpenCode, Hermes, and PI. The figures come from six professional work benchmarks, including OfficeQA with 529 real-world queries.
- MarkTechPostNewsNace AI releases Drex for Search, a retrieval layer for AI agents
AINace AI has released Drex for Search, a hosted retrieval system for AI agents that narrows documents by folder, file, section and page before passing only relevant pages to an answering model. Nace reports 92.05% accuracy on BrowseComp-Plus with GPT-5 using 12.61 search calls per question, and 77.5% R@1 on VisDocAgentBench. The system is accessible through a CLI, Python and TypeScript SDKs, MCP, and REST, and it can run on Nace's cloud, a dedicated cloud, or on-prem. All benchmark figures are as published by Nace AI.
- 🚨 AI News | TestingCatalogXNace AI launches Drex for Search to help agents retrieve exact file pages
AINace AI launched Drex for Search, a tool that lets agents find answers in cloud-stored files by locating the folder, file, and page, then returning the answer, quote, and page number. Nace reports 62% fewer steps and 56% fewer input tokens per question, and agents answered 67.1% of 529 queries correctly across six work benchmarks, up from 63.1%.
