prompt logo

prompt

Free

Design an open-source deep research agent system

FreeFree tier
Inputs: textOutputs: text
Type
Open Source

About prompt

This prompt, named 'Open Deep Research Agent Architect', instructs an AI to design a comprehensive open-source deep research agent system capable of competing with commercial products like OpenAI Deep Research, Gemini Deep Research, and Perplexity Pro. It provides a detailed blueprint covering data pipeline, training recipe, inference modes, tool stack, evaluation harness, deployment topology, and governance. The system is designed to handle multi-hop, evidence-bound questions over the open web with verifiable citations, long-horizon planning, and reproducible runs. Key components include planning a research trajectory across 20–40+ turns, searching/browsing the web, running code, tracking evidence as a typed graph, detecting contradictions using independent sources, synthesizing with citations, and logging every action for reproducibility. The prompt also defines a task contract with input/output shapes, citation policies, and a refusal policy for unanswerable questions.

Key Features

Plans a research trajectory across 20–40+ turns
Searches the web, browses pages, runs code, reads documents
Tracks evidence as a typed graph
Detects contradictions and triangulates multiple independent sources
Synthesizes with verifiable citations
Logs every action for fully reproducible runs
Defines a task contract with input/output shape, deadline, and citation policy
Includes a refusal policy for unanswerable questions

Pros & Cons

Pros
  • Open-source and free to use
  • Provides a detailed, structured blueprint for building research agents
  • Emphasizes reproducibility and verifiable citations
  • Designed to handle complex, multi-hop questions over the open web
  • Includes mechanisms for contradiction detection and source triangulation
Cons
  • Only a design prompt, not a ready-to-use implementation
  • Requires significant technical expertise to realize the full system
  • The output is a specification, which may not be immediately deployable

Best For

Designing open-source deep research agent systemsCreating AI agents that compete with commercial research toolsBuilding multi-hop evidence-based answer systems for complex questionsDeveloping reproducible research pipelines with citation tracking

FAQ

What is the Open Deep Research Agent Architect prompt?
It is a detailed system prompt designed to instruct an AI to generate a comprehensive architecture for an open-source deep research agent, covering data pipeline, training, inference, and governance.
How does this prompt differ from standard RAG prompts?
This prompt goes beyond simple retrieval-augmented generation by requiring multi-step planning, evidence graph construction, contradiction detection, and full action logging for reproducibility.
Is the prompt free to use?
Yes, the prompt is hosted on GitHub in a public repository (ai-boost/awesome-prompts) and is free and open-source.