How it works

Audience: End users

How it works

GL Open DeepResearch follows a simple, predictable flow from your question to a structured answer.

Simple flow

Step-by-step

1. You send a question — Submit a natural-language research query (e.g., "What are the main arguments for and against renewable energy?").

2. Choose a profile — Select how deep or fast you want the research:

   • Quick profiles — Faster answers, fewer steps (e.g., ESSENTIAL)

   • Deep profiles — More thorough research, more sources, longer time (e.g., COMPREHENSIVE, INTERNAL)

3. The service processes — GL Open DeepResearch:

   • Selects the right research engine based on your profile

   • Breaks down your question into sub-questions

   • Gathers information from multiple sources using search and retrieval tools

   • Synthesizes findings into a coherent answer

4. You receive results — Get back:

   • A structured answer to your question

   • Sources and references (depending on the API response)

   • Optional thinking steps and tool usage (if using streaming)

What happens behind the scenes

• No configuration needed — The service handles research engine selection, model configuration, and tool setup based on your profile.

• Automatic source gathering — The research engine searches, retrieves, and evaluates multiple sources automatically.

• Evidence-based synthesis — Answers are built from gathered sources, not generated from memory.

The entire process is handled by the service—you just send a question and get a researched answer.

For examples of questions and their results with step-by-step instructions, see Examples. For when to use it, see When should I use it?.