Personalized configuration system for report generation
Open Deep Research offers comprehensive configuration options to meet different research needs. In terms of report structure, the system supports the use of both standard research report formats and allows users to fully customize the section layout via the report_structure parameter. In terms of research depth control, users can set the number of search iterations per section via the max_search_depth parameter (default 2), and regulate the number of search queries generated per section via the number_of_queries parameter (default 2).
The tool also has a powerful model selection function: openai or groq can be selected as the provider in the planning stage, and specific models support o3-mini, etc.; the writing stage supports anthropic, openai, and other providers, and the latest version of Claude 3.5 Sonnet is used by default; and the search API supports the two mainstream services of Tavily and Perplexity two mainstream services. All these parameters can be dynamically adjusted through configuration files or runtime parameters.
Of particular note, the system is designed with an interactive feedback mechanism that allows the user to request adjustments (e.g., "Include a revenue estimate") through natural language commands after generating the initial report plan, and final confirmation before proceeding to the formal research and writing phase.
This answer comes from the articleOpen Deep Research: LangChain's Open Source Intelligent Assistant for Deep ResearchThe
English 
简体中文 
日本語 
Deutsch 
Português do Brasil