AI researches. Humans verify. Knowledge grows. Open for all.
Every piece of knowledge is researched by AI, verified by humans, and connected in a graph that grows every day.
Every result passes community peer review. Minimum 25 verified reviewers. No result enters the knowledge graph without human approval.
Five specialized AI agents refine each result through up to 20 rounds. Only results scoring 0.98 or higher pass the quality gate.
A connected graph of verified science. Searchable, explorable, and queryable through a free public API. Knowledge as a public good.
Contribute CPU, GPU, or API access by running the open-source worker. Earn points, badges, and recognition on the leaderboard.
Coordination detection, behavioral fingerprinting, and automatic escalation for sensitive topics. Science stays honest.
Knowledge is re-verified periodically. Outdated results are flagged. Version history tracks how understanding evolves over time.
Every piece of knowledge follows the same rigorous path. No exceptions.
Runs 24/7 on a secure server in Germany. Autonomously plans scientific research across all disciplines. Identifies knowledge gaps and prioritizes them.
Retriever, Planner, Stylist, Visualizer, and a Multi-Critic ensemble. Up to 20 iterative refinement rounds. Only scores of 0.98 or higher pass.
Minimum 25 verified human reviewers. Trust-weighted voting. Honeypot calibration. Lobby and coordination detection running silently in the background.
Verified results become connected nodes. Cross-domain links are detected automatically. Gaps in knowledge drive future research direction.
Search, explore, and query verified knowledge for free. Developers build on the public API. The Knowledge Guide answers questions from verified sources only.
Everything you need to contribute to and benefit from OpenGen is open source. The core stays protected.
| Component | License | What it does |
|---|---|---|
| Worker client | Open source | Contribute compute — runs on your machine |
| Data models | Open source | Shared format between worker and server |
| API client | Open source | Query the verified knowledge graph |
| Documentation | Open source | Architecture, verification, decisions — all public |
| Master Agent | Closed | Research planning and orchestration |
| Pipeline agents | Closed | 5-agent refinement engine |
| Knowledge graph | Closed | Storage, connections, gap detection |
Core infrastructure. LLM client, pipeline agents, knowledge graph engine. Building the foundation.
Full 5-agent pipeline with dynamic refinement loop. Research scheduler. REST API.
This website becomes a live dashboard. Accounts, leaderboard, community review interface.
Worker network goes live. Open-source client, security layer, distributed task execution.
Public knowledge API. Multilingual support. Knowledge Guide. Custom model training.
OpenGen is self-funded by two founders from Germany. We have the vision and the plan. Every contribution - no matter how small - directly funds infrastructure and research capacity. There are no salaries. Every euro goes into building.
Fund open scientific infrastructure. Your name on the supporters page.
Any amount helps. Covers server costs, API calls, and development time.
Share our values of ethical, open AI? Let's talk partnership.