Hugging Face
The leading agent for drug discovery. Axon helps modern drug discovery teams move faster than ever before.
01
Novel insights at blinding speed
Cover ground faster than ever and surface insights that would have stayed hidden.
02
Your tools, all in one place
Slots into your existing workflow — your tools, your data, your infrastructure. Ready in minutes.
03
Agents that scale and grow
Handles the heavy lifting across your pipeline and gets sharper with every interaction.
Do more with Axon
See more examples of Axon at work
Frequently asked questions
Axon is an agent harness for preclinical drug discovery. It helps scientists run computational tools, analyze results, generate reports, and coordinate multi-step discovery work from a single interface.
Axon is built for teams working across preclinical discovery, including medicinal chemists, computational chemists, CADD scientists, DMPK / PK-PD teams, structural biologists, cheminformaticians, and discovery program leads.
Axon can help with tasks such as conformer generation, docking preparation, virtual screening, molecular dynamics setup, ADMET analysis, structure review, literature-grounded research, tool execution, and report generation. The exact capabilities depend on the tools and data available in your deployment.
General-purpose chat models are useful for discussion and drafting. Axon is designed to do scientific work: it can call domain tools, run code, inspect outputs, use discovery-specific context, and produce auditable artifacts that scientists can review.
No. Axon is designed to make computational methods more accessible to scientists who know the biology, chemistry, and program context. Expert users can still go deeper, inspect intermediate steps, and guide the system when needed.
No. Axon is a force multiplier for scientific judgment. It helps scientists move faster through setup, execution, analysis, and reporting, while keeping domain experts in control of decisions.
Yes. In enterprise deployments, Axon can be configured to use your internal tools, data sources, compute environments, and reporting formats. The goal is to make Axon fit the way your team already works, rather than forcing a new discovery stack.
Yes. Enterprise deployments can run in private cloud or customer-controlled infrastructure, depending on your security and integration requirements.
Your proprietary targets, molecules, assays, programs, and internal data remain under your control. Enterprise deployments can be configured so Axon does not retain or train on proprietary scientific data or IP.
Yes. Axon records the steps it takes, including tool calls, code execution, intermediate outputs, and final results. Scientists can inspect, interrupt, redirect, or rerun work instead of treating the output as a black box.
Axon grounds its work in executable tools, retrieved context, generated artifacts, and reviewable traces. Like any AI system, it can make mistakes, but its work is designed to be inspectable so scientists can see how a result was produced.
Most teams start by identifying one concrete discovery workflow or bottleneck, then testing Axon on public data or a controlled evaluation environment. From there, Axon can be expanded into private tools, proprietary data, and enterprise deployment.
