Open Deep Search: Building Transparent, Open-Source Search-Augmented LLMs

In this episode we dive into Open Deep Search (ODS), the open-source path to turning LLMs into smart, real-time researchers. We break down the Open Search Tool and Open Reasoning Agent, explain how query rephrasing, structured results, and embeddings improve accuracy, and discuss why openness matters for transparency and collaboration. We also compare ODS’s V1 (React with CoTE/CoTSC) and V2, and surface how ODS stacks up against big-name search AIs on benchmarks like SimpleQA and Frames.

Note: This podcast was AI-generated, and sometimes AI can make mistakes. Please double-check any critical information.

Sponsored by Embersilk LLC