An interactive portal into open, agentic science

Scientific knowledge is expanding at an exponential pace, yet accessing and navigating it still feels like walking through a maze of PDFs, search bars, and closed networks. At BeeARD, we believe science shouldn’t feel like a gated archive, it should feel like an open, explorable world.

That’s why we built GraphSurfer.

GraphSurfer is a web-based, interactive platform that lets anyone explore scientific knowledge graphs in a natural, visual, and playful way. Think of it as a map of scientific thought, where each node represents a concept, paper, or entity, and each edge reflects a meaningful connection between them.

Instead of searching and reading one paper at a time, you “surf” through a network of ideas, zooming out to see the big picture or diving into the details of a specific research cluster. Whether you’re a scientist, student, or curious generalist, GraphSurfer offers a new way to engage with science: intuitive, dynamic, and curiosity-driven.

At BeeARD, we’re developing the infrastructure for agentic science: a new approach to scientific discovery powered by multi-agent AI systems and dynamic knowledge graphs. At the heart of this system is ARD, our Autonomous Research Discovery framework. ARD connects reasoning agents to structured knowledge representations, allowing for large-scale, automated hypothesis generation and research synthesis.

But even the smartest AI agents are only as impactful as the interfaces people use to interact with them. While we were building HypGen, our AI-driven hypothesis generation platform, we realized something crucial:

Most people, even scientists, don’t have intuitive tools for navigating the kinds of knowledge graphs we use to generate hypotheses.

We needed a way for real humans to see what our agents see. We wanted to give people a way to explore, question, and contribute - not just receive outputs.

GraphSurfer was born out of that need.

GraphSurfer connects directly to the knowledge graphs powering our hypothesis generation systems. In this first version, we’ve launched a graph focused on psychedelic science: a rich, rapidly evolving field full of interconnected ideas, emergent subfields, and gaps in the literature.

• Knowledge Graph Interface: A visual, clickable graph of interconnected scientific concepts and entities. Users can click nodes to expand related concepts, learn about them, or jump between areas of research.

• Kai, the AI Surf Guide: Kai is our conversational assistant embedded into GraphSurfer. He explains concepts, shows how they’re connected, and guides you through the research space in natural language.

• User-Driven Discovery Loops: As users explore the graph, their navigation patterns (what they click, what they ask about, what paths they follow) are logged and analyzed - not for surveillance, but for scientific creativity. These interactions surface interesting subgraphs, areas where human curiosity clusters, which are then used to seed future hypotheses in HypGen.

Here’s where it gets interesting.

GraphSurfer doesn’t just visualize knowledge. It’s actively feeding the discovery process. When people explore the graph, they help us identify which parts of the scientific landscape are attracting attention. These attention clusters, what we call curiosity-driven subgraphs, are sent as prompts to HypGen, our autonomous hypothesis generator.

HypGen then runs multi-agent workflows to generate novel, scientifically grounded hypotheses based on those prompts. The hypotheses are published to hypgen.ai, where they can be discussed, validated, and iterated on by the research community.

In other words:

• GraphSurfer = Human curiosity surfaces interesting problems

• HypGen = AI agents turn those problems into testable hypotheses

• Researchers = Refine, explore, and experiment with the outputs

This is a feedback loop between human and machine intelligence, driven by exploration, not just extraction.

GraphSurfer and HypGen are both fully open-source. We believe that the future of science isn’t just faster or smarter, it’s more participatory, more accessible, and more transparent.

Any lab can approach us to build a knowledge graph around their specific research domain. We use their papers and data to create a structured representation of their field. Once live, that graph becomes explorable in GraphSurfer and generative in HypGen, tailored to their expertise, but open to global contributors.

Whether you’re a biologist, a funder, or a curious outsider, this model lets you interact with science in a way that’s never been possible before.

We’re actively developing new features for GraphSurfer, including:

• Support for multiple knowledge domains (not just psychedelics)

• Token-based incentives for exploration, contribution, and curation

• Leaderboards and achievement systems for scientific discovery

• Deeper Kai integrations to support contextual Q&A, summarization, and guidance

We’re also working closely with communities like PsyDAO to customize GraphSurfer for specific research areas, including new tools to support public science funding, token-gated discovery, and collaborative graph-building. PsyDAO recently awarded a 5 ETH grant to support the development of GraphSurfer, helping us expand the psychedelic knowledge graph and improve the platform’s usability, gamification, and community features for psychedelic science explorers.

v1:

Science is too important to be locked behind paywalls, institutional silos, or PDF stacks. At BeeARD, we’re building an ecosystem that opens it up: not just for scientists, but for everyone.

GraphSurfer is one of the doors into that ecosystem. A portal for exploring the known, discovering the unknown, and contributing to the future of knowledge itself.

We invite you to surf the graph, follow your curiosity, and help shape the next wave of open science.

Explore now: https://graphsurfer.xyz [pw: psychonaut2024]

Try HypGen: https://hypgen.ai [Stay tuned! In the next blogpost we dive deep into HypGen.]

Build with us: info@beeard.ai

An interactive portal into open, agentic science

Scientific knowledge is expanding at an exponential pace, yet accessing and navigating it still feels like walking through a maze of PDFs, search bars, and closed networks. At BeeARD, we believe science shouldn’t feel like a gated archive, it should feel like an open, explorable world.

That’s why we built GraphSurfer.

GraphSurfer is a web-based, interactive platform that lets anyone explore scientific knowledge graphs in a natural, visual, and playful way. Think of it as a map of scientific thought, where each node represents a concept, paper, or entity, and each edge reflects a meaningful connection between them.

Instead of searching and reading one paper at a time, you “surf” through a network of ideas, zooming out to see the big picture or diving into the details of a specific research cluster. Whether you’re a scientist, student, or curious generalist, GraphSurfer offers a new way to engage with science: intuitive, dynamic, and curiosity-driven.

At BeeARD, we’re developing the infrastructure for agentic science: a new approach to scientific discovery powered by multi-agent AI systems and dynamic knowledge graphs. At the heart of this system is ARD, our Autonomous Research Discovery framework. ARD connects reasoning agents to structured knowledge representations, allowing for large-scale, automated hypothesis generation and research synthesis.

But even the smartest AI agents are only as impactful as the interfaces people use to interact with them. While we were building HypGen, our AI-driven hypothesis generation platform, we realized something crucial:

Most people, even scientists, don’t have intuitive tools for navigating the kinds of knowledge graphs we use to generate hypotheses.

We needed a way for real humans to see what our agents see. We wanted to give people a way to explore, question, and contribute - not just receive outputs.

GraphSurfer was born out of that need.

GraphSurfer connects directly to the knowledge graphs powering our hypothesis generation systems. In this first version, we’ve launched a graph focused on psychedelic science: a rich, rapidly evolving field full of interconnected ideas, emergent subfields, and gaps in the literature.

• Knowledge Graph Interface: A visual, clickable graph of interconnected scientific concepts and entities. Users can click nodes to expand related concepts, learn about them, or jump between areas of research.

• Kai, the AI Surf Guide: Kai is our conversational assistant embedded into GraphSurfer. He explains concepts, shows how they’re connected, and guides you through the research space in natural language.

• User-Driven Discovery Loops: As users explore the graph, their navigation patterns (what they click, what they ask about, what paths they follow) are logged and analyzed - not for surveillance, but for scientific creativity. These interactions surface interesting subgraphs, areas where human curiosity clusters, which are then used to seed future hypotheses in HypGen.

Here’s where it gets interesting.

GraphSurfer doesn’t just visualize knowledge. It’s actively feeding the discovery process. When people explore the graph, they help us identify which parts of the scientific landscape are attracting attention. These attention clusters, what we call curiosity-driven subgraphs, are sent as prompts to HypGen, our autonomous hypothesis generator.

HypGen then runs multi-agent workflows to generate novel, scientifically grounded hypotheses based on those prompts. The hypotheses are published to hypgen.ai, where they can be discussed, validated, and iterated on by the research community.

In other words:

• GraphSurfer = Human curiosity surfaces interesting problems

• HypGen = AI agents turn those problems into testable hypotheses

• Researchers = Refine, explore, and experiment with the outputs

This is a feedback loop between human and machine intelligence, driven by exploration, not just extraction.

GraphSurfer and HypGen are both fully open-source. We believe that the future of science isn’t just faster or smarter, it’s more participatory, more accessible, and more transparent.

Any lab can approach us to build a knowledge graph around their specific research domain. We use their papers and data to create a structured representation of their field. Once live, that graph becomes explorable in GraphSurfer and generative in HypGen, tailored to their expertise, but open to global contributors.

Whether you’re a biologist, a funder, or a curious outsider, this model lets you interact with science in a way that’s never been possible before.

We’re actively developing new features for GraphSurfer, including:

• Support for multiple knowledge domains (not just psychedelics)

• Token-based incentives for exploration, contribution, and curation

• Leaderboards and achievement systems for scientific discovery

• Deeper Kai integrations to support contextual Q&A, summarization, and guidance

We’re also working closely with communities like PsyDAO to customize GraphSurfer for specific research areas, including new tools to support public science funding, token-gated discovery, and collaborative graph-building. PsyDAO recently awarded a 5 ETH grant to support the development of GraphSurfer, helping us expand the psychedelic knowledge graph and improve the platform’s usability, gamification, and community features for psychedelic science explorers.

v1:

Science is too important to be locked behind paywalls, institutional silos, or PDF stacks. At BeeARD, we’re building an ecosystem that opens it up: not just for scientists, but for everyone.

GraphSurfer is one of the doors into that ecosystem. A portal for exploring the known, discovering the unknown, and contributing to the future of knowledge itself.

We invite you to surf the graph, follow your curiosity, and help shape the next wave of open science.

Explore now: https://graphsurfer.xyz [pw: psychonaut2024]

Try HypGen: https://hypgen.ai [Stay tuned! In the next blogpost we dive deep into HypGen.]

Build with us: info@beeard.ai