For Drug Discovery Teams

Dock Molecules, Score ADMET, Generate Leads — All in One Place

Run DiffDock, AutoDock Vina, ADMET-AI, and REINVENT4 without installing anything. From target to lead candidate in minutes. Free for researchers.

No install · No GPU · Free for .edu emails (750 credits/month)

The Drug Discovery Pipeline

Four steps from target to lead — each running on cloud GPUs.

1

Identify Target

Know your protein target structure

ESMFold / Boltz-2
~3-30s
Pocket Detection
~2s
2

Dock

Dock small molecules to the target

DiffDockAI-powered
~15s
AutoDock Vina
~10s
Chai-1Complex
~30s
3

Score ADMET

Predict drug-likeness and toxicity

ADMET-AI41 endpoints
~1s
Molecular Properties
instant
4

Generate

Design novel molecules de novo

REINVENT4
~20s
SA Score
~1s

DiffDock vs AutoDock Vina vs Chai-1

Three docking approaches for different use cases. All available on SciRouter.

FeatureDiffDockAutoDock VinaChai-1
MethodDiffusion model (AI)Physics-based scoringCo-folding (AI)
Speed~15 seconds~10 seconds~30 seconds
Requires binding site?No (blind docking)Yes (grid box)No
Ligand flexibilityFull (SE(3) diffusion)Rotatable bondsFull
Confidence scoreYes (ranked poses)Binding energy (kcal/mol)pLDDT + pTM
Protein-ligand complex?YesYesYes + DNA/RNA
Cost on SciRouter10 credits ($0.10)5 credits ($0.05)15 credits ($0.15)
Best forBlind screening, novel targetsKnown pockets, fast iterationMulti-component complexes

Frequently Asked Questions

Start docking molecules in 30 seconds

Paste a SMILES string, pick a target, see docking poses. No GPU, no Docker, no waiting. Free for researchers.