Molecular dynamics, in your workflow
Use Claude or ChatGPT through Codex to prepare structures, run simulations on SimAtomic’s remote GPUs, and analyze results in your local project.
Your assistant guides the workflow. SimAtomic runs the compute.
You set the questionBring your structure and review the plan.
Claude or Codex coordinatesPrepare, submit, monitor, and analyze.
SimAtomic runs the simulationRemote GPU compute and downloadable results.
Research community
Research support has contributed to a cardiac troponin study in ACS Omega and SimAtomic participates in the Scaleway Startup Program. View the publication ↗
A workflow you can follow
Keep the conversation, input files, analysis scripts, and outputs together in a local project. The current preparation workflow supports soluble proteins, optionally with one prepared small molecule. Explore the scientific capabilities →
Ask your assistant to inspect the structure, identify preparation issues, and propose settings for a short test.
Prompt: Check this structure and plan a short MD test. Explain any input issues and wait for my approval before uploading or submitting.
Review the inputs, settings, and estimated credits when available. Approve the run, then ask your assistant to follow its progress. Multiple jobs can run in parallel.
No local GPU is needed for the simulation.
Download trajectories and output files. Ask your assistant to write analysis scripts, produce plots, and explain what the results can—and cannot—tell you.
Review the methods and reuse the saved scripts.
See the complete example
Follow a condensed Claude session from an input issue through structure preparation, an approved submission, job monitoring, and analysis of a completed run.
The interactive demo pairs the conversation with a real trajectory viewer, including frame controls, structural metrics, and molecular contacts.
Explore the chat + trajectory demo →The example is a 5 ps workflow test. Production timescales and replicates should match your scientific question.
Prepared inputsStructures and the settings used for the run
Simulation outputsDownloaded trajectory and result files
Reusable analysisScripts and figures saved by your assistant
Bring the assistant you already use
Connect SimAtomic once, authorize access, and work from a local project folder.
Add the SimAtomic connector, connect a local folder, and ask Claude to inspect the available tools.
Set up Claude →See the example sessionAdd SimAtomic in the desktop settings and authorize with your API key. Use Codex to work with local inputs, downloads, and scripts.
Set up Codex →This connects the local Codex workflow; ChatGPT on the web has a separate integration setup.
Submit jobs, monitor progress, and download results from your own scripts.
You can also work directly in the browser with SimAtomic’s simulation and analysis interface.
Pricing & billing
SimAtomic credits pay for jobs submitted to the SimAtomic API, including jobs started through the Claude or Codex connector. Standard compute uses one credit per minute.
| Package | Price | Credits | Estimated compute | Rate | Best for |
|---|---|---|---|---|---|
| Starter | $120 | 1,000 | ~16 hours | $0.12 / credit | Students and PIs evaluating the platform |
| Connector choicePro | $500 | 5,000 | ~83 hours | $0.10 / credit | Academic labs and biotech startups running weekly jobs |
| Team | $960 | 12,000 | ~200 hours | $0.08 / credit | Small biotech teams running many simulations |
| Enterprise | Custom | Volume | Custom | Volume pricing | Organizations needing custom integrations, support, or deployment |
Starter: UI access, run metadata exports, and email support. Pro: API + UI access and priority support. Team: multiple seats, shared billing and spend controls, a dedicated support contact, and onboarding. Enterprise options include custom integrations, audit logs, SSO, and deployment options.
The Claude and Codex connectors require SimAtomic API access. The current pricing defines API + UI access on Pro.
Your assistant subscription, usage limits, and any additional provider charges are separate from SimAtomic credits. If you use an OpenAI or Anthropic API key for the assistant, that provider’s API billing applies.
Your SimAtomic API key and compute credits are separate from your assistant account.
Plan your first run
On SimAtomic’s remote modern GPUs. Your local project holds inputs, downloaded results, and analysis scripts. Local analysis uses the tools installed on your computer.
Ask your assistant to show the proposed inputs and settings, and wait for your approval before uploading or submitting. The setup guides include prompts for this workflow.
Jobs can run in parallel. Compute jobs require available credits. Contact SimAtomic about dedicated GPU capacity for your team.
Inspect the preparation, simulation settings, and analysis. A short test checks the workflow; scientific conclusions need appropriate timescales, controls, and replicates.