Build chemical software.
Keep the evidence.
From a first Python script to a reproducible scientific pipeline. Run calculations, test ideas and turn results into a network you can investigate.
ct-cli run hello.py# Send the same script to a workerct-cli submit hello.py --wait One job system, end to end
Compose ordinary jobs. Reuse accepted results.
What are you working on?
Guides for everyday scientific workOptimize molecular structures
Keep the final geometry, convergence and trajectory together.
Build reaction networks
Connect sourced proposals, observations and calculation evidence.
Run molecular simulations
Submit an allocation. Parse trajectories. Save the observations.
Compare algorithms
Freeze a dataset and measure candidates with the same scoring rules.
Evaluate DFT energies
Run fixed-geometry calculations on CPU or GPU with explicit provenance.
Improve throughput
Measure useful work, setup costs and failures before tuning capacity.
From script to shared workflow
Use the simplest unit that fits. Add a typed contract when you need reuse, dependencies when you need a pipeline, and scoring when you need a benchmark.
Bring your code. Choose where it runs.
Use a registered CPU, GPU or Ray worker. Keep source, runtime and scientific evidence explicit as your workload grows.