YEARS
Discovering a single new material can take years and cost millions.
Early Stage - Building in public
The IDE for discovering new materials
Set an objective. Generate and test compounds. Simulate before you ever touch a lab bench. An AI-native environment for materials science, built for scientists and powered by specialized models and agents.
objective -> generate -> simulate -> lab -> repeat
Materials discovery is stuck in the past
Discovering a single new material can take years and cost millions. Scientists switch between disconnected tools: literature databases, DFT and simulation software, spreadsheets, and lab notebooks. There is no environment where AI can meaningfully assist the entire loop, from idea to validated compound.
DISCONNECTED TOOLS
Literature databases, DFT and simulation software, spreadsheets, and lab notebooks do not naturally work together.
NO AI-NATIVE WORKFLOW
No shared environment lets AI assist the full path from objective to validated compound.
The Core Loop
One loop, from idea to lab
01 - OBJECTIVE
Set the Objective
Define what you are looking for: target properties, constraints, application. Plain language or structured spec.
Plug in the intelligence you trust
Material Genesis does not lock you into one model. Connect whatever reasoning system fits the job.
Materials-Science Models
Connect models trained specifically on materials science: property prediction, structure generation, and domain-specific reasoning.
General-Purpose LLMs
Bring your own LLM for reasoning, literature synthesis, and hypothesis generation alongside domain models.
Autonomous Agents
Deploy agents that run parts of the discovery loop independently, proposing candidates, running simulations, and reporting back.
Collaborate like it is one lab, not five tools
Invite collaborators, share objectives and results in real time, and keep your whole team, human and AI, working from the same discovery loop.
JKMSAL+3
Concept visual: shared workspace, product in development.
Work the way you already work
IDE (Desktop)
A dedicated environment for deep, focused discovery work: your objectives, candidates, simulations, and results in one place.
CLI
Script and automate discovery pipelines. Integrate into existing computational workflows and CI-style research loops.
Web
A lightweight, accessible interface for quick objectives, collaboration, and reviewing results from anywhere.
Why This, Why Now
The pieces just came together
AI models are finally good enough at chemistry and materials reasoning. Simulation is cheap enough to run in a loop. The missing piece is the environment that ties objective-setting, generation, simulation, and lab handoff together, the same way code editors evolved once LLMs got good enough to pair-program.
Are you a scientist or researcher? Help us build this.
We are in early development and looking for scientists, researchers, and labs to help shape the product from the ground up. Your domain expertise directly shapes what we build. Tell us about yourself and we will be in touch.