Seeking Collaborators & Funding! We're actively looking for contributors and funding to accelerate development. Contact us at a.siddiki@proton.me.
Build Advanced AI Research Systems, Faster.
Leverage the Open Deep Research API framework: A modular, extensible foundation for creating specialized, multi-agent AI research capabilities.
Powered by Luminary AI Solutions LLC
Stop Reinventing the Wheel for AI Research
Building robust AI research systems is complex. ODR-API provides the structured, reusable foundation you need to focus on innovation, not infrastructure.
The Challenge
Building sophisticated research agents often means:
- Duplicated effort for common tasks (search, scraping).
- Complex agent coordination and data flow.
- Difficulty adapting to new domains or tools.
- Scalability and maintainability issues.
The ODR-API Solution
Our framework provides:
- Modular `Agencies` for domain specialization.
- Reusable `Services` for core functionalities.
- Clear `Agent` orchestration patterns.
- Reliable data flow with `Pydantic`.
Powerful Features Out-of-the-Box
ODR-API provides a comprehensive toolkit for building sophisticated AI research applications.
Create specialized research domains as independent 'Agencies', each orchestrating their own workflow and agents for specific tasks (e.g., finance, legal, deep web).
Leverage a shared library of powerful components for search, advanced scraping (Crawl4AI), chunking, PDF handling, ranking, and more across any agency.
Coordinate multiple specialized LLM agents (like Planner, Writer, Refiner) within each agency using customizable workflow logic for complex research processes.
Ensure reliable data transfer and validation between agents and services using Pydantic V2 schemas, enabling robust and predictable interactions.
Built for growth. Easily add new agencies, integrate custom LLM tools, swap out services, or configure LLM providers and parameters via environment or API.
Monitor the research process live with detailed step-by-step updates streamed via WebSockets, providing transparency into the agent's execution.
Built on a Solid, Modular Foundation
Understand the core components that make ODR-API powerful and easy to extend.
Core Architecture Flow
(e.g., Calculator, API Call)
Each Agency (`app/agencies/...`) defines a specific research capability (e.g., deep research, financial analysis). It bundles its own orchestration logic (`orchestrator.py`), specialized LLM agents (`agents.py`), and data structures (`schemas.py`).
This modular approach allows focused development and easy addition of new research domains without impacting others.
Pydantic Schemas: The Data Glue
Pydantic models define the data contracts between all components (Agencies, Agents, Services, Tools, API), ensuring reliable, validated communication and structured LLM interactions.
How the Deep Research Agency Works
See the orchestrated flow of agents and services in action within the example `deep_research` agency.
1. Define & Plan
Provide a research query. The Planner agent generates a structured plan, including search queries and an outline.
2. Gather & Process
The system executes searches, scrapes web content (Crawl4AI), handles PDFs, chunks text, and ranks information for relevance (Together AI).
3. Synthesize & Write
The Writer agent drafts the report using the processed context, structuring it according to the plan and citing sources.
4. Refine & Iterate
The Refiner agent iteratively enhances the draft, incorporating only new information from further searches if needed, ensuring comprehensive coverage.
5. Finalize & Deliver
The system assembles the final report, formats citations, adds references, and delivers it via WebSocket stream.
Built with a Modern, Robust Stack
Leveraging powerful libraries and services for optimal performance and developer experience.
Backend Framework
AI & Core Services
Frontend Demo
Ready to Build Your AI Research System?
Experience the power and flexibility of ODR-API. Try the interactive demo or dive straight into the source code.