01
Protein structure prediction
The first planned capability: protein folding and complex prediction, with downloadable structures and confidence information.
RESEARCH PILOT · APPROVAL REQUIRED
Protein structure prediction.
A clearer path
to discovery.
From sequence to research insight. A planned workspace for molecular predictions, confidence review and reproducible computational biology workflows.
Execution is not yet available. No prediction charges or biological submissions are enabled.
Big possibilities.AI-generated illustration · not a measured structure
01 Research-led02 Client approval03 Scientific provenance04 Budget control
THE RESEARCH WORKFLOW
Ask better molecular questions.
02
Protein design
Future binder-design workflows to generate and evaluate candidates. Laboratory validation remains essential.
03
Protein embeddings
Future sequence representations for research analysis and downstream bioinformatics workflows.
HUMAN JUDGMENT. COMPUTATIONAL REACH.
A better starting point.
A scientist at the center.
Move from a research question to a structured computational experiment. Review model confidence, preserve provenance and decide what deserves laboratory validation.
Find your research workflow ↗01 · Define the question02 · Review the prediction03 · Validate experimentally
BUILT AROUND YOUR RESEARCH
For the teams behind the next question.
Academic labs and core facilities
Explore AI protein structure prediction for computational biology and structural biology research. Review confidence information and validate findings experimentally.
Biotech and AI-native drug discovery
Evaluate protein complexes and molecular hypotheses with traceable model versions, project controls and a path to reproducible workflows.
Pharmaceutical research teams
Scope computational protein research around access controls, documented processing regions, provider review and project-level budgets.
Contract research organizations
Plan structured bioinformatics workflows across approved client projects, with workspace separation and a reviewable job history.
Research platforms and developers
Explore a future protein prediction API integration with job tracking, budget controls and downloadable scientific artifacts. Third-party access requires written approvals.
CONTROL AT EVERY STEP
Your project. A traceable workflow.
Private workspaces
Planned tenant-level permissions, controlled artifact access and explicit retention settings.
Budget before compute
Estimates, spending ceilings and Treasury reconciliation are release requirements. Retail rates are not yet finalized.
Transparent execution
Available providers, model licenses and processing regions will be disclosed before any submission. Japan-hosted execution is not EU-only.
START WITH A RESEARCH BRIEF
Good science starts with a clear scope.
Prepare your organization, research objective and expected workload. AGICY must approve each client and use case before enabling execution. Resale, third-party API access and material changes of scope require separate written approval.
Prepare pilot brief →Read draft research service terms ↗Before you begin
Is AGICY Bio available now?
This is a pilot preview. Provider approval, client approval and technical validation are required before execution becomes available.
Are predictions experimentally validated?
A model prediction is a research hypothesis. Confidence metrics are not clinical success probabilities, and laboratory validation is your responsibility.
Which models will be offered?
Boltz2 is the initial evaluation target. The enabled catalogue, model versions, licenses and limits will be confirmed before activation.
Can I integrate it into my own product?
Developer access is planned. Serving third parties or reselling access requires AGICY's express prior written approval and any applicable upstream permission.