deep-research

Conduct fact-checked research by orchestrating iterative web searches, parallel synthesis, and dual-reviewer claim verification over external sources. This protocol validates high-stakes claims through a critic challenge mechanism to ensure evidence-grounded accuracy.

Dual-reviewer verification validates research claims by using two independent reviewers to cross-check evidence gathered from iterative web searches. This mechanism challenges high-stakes claims to mitigate bias and ensure factual accuracy before generating a cited report.

Yes, you can use external search APIs to gather evidence iteratively. The research protocol dispatches parallel synthesis workers to process the retrieved external sources, grounding the final answers and cited report in verified evidence.

Yes, external research tools and search APIs are required dependencies. The parallel synthesis workers rely on these external tools to retrieve and process evidence iteratively, ensuring the dual-reviewer verification protocol has sufficient data to validate high-stakes claims.

The best way to generate a cited report is through an orchestrated research protocol combining iterative web searches, parallel synthesis, and dual-reviewer verification. This approach cross-validates high-stakes claims through a critic challenge before producing the final cited document.

Avoid using this dual-reviewer research protocol when external search APIs and research tools are unavailable, or when tasks do not require high-stakes claim verification. The iterative search and parallel synthesis overhead makes it unsuitable for simple, single-query lookups.