Introduction
Run molecular modeling jobs with the Boltz API.
The Boltz API lets you run structure prediction, molecular design, and library screening from code.
Who these docs are for
Section titled “Who these docs are for”Each guide shows examples for the audience that matches how you work:
- Scientist — you want to run jobs from your own machine, have results land in a local directory, and resume long-running work after a restart. Use the Python
Boltzclient’s higher-levelrun_*/start_*methods, or theboltz-apiCLI with a YAML input file. - Platform Developer — you’re integrating Boltz into a service or app and need direct control over the request/response lifecycle. Use the Python or Typescript SDK to start jobs, poll status, and read download URLs yourself.
- Agent — you want Claude Code, Codex, Cursor, or another coding agent to run Boltz jobs for you. Use the CLI-backed agent plugins/skills when available, or give the agent the permission-friendly
boltz-apicommand pattern.
Start with a job
Section titled “Start with a job”- Predict structure and binding — Submit a molecular complex and get predicted structures back.
Predict structure and binding → - Design novel molecules — Generate new small molecules or protein binders with streaming results and optional early stopping.
Small molecule design → · Protein design → - Screen a library — Score your own small molecules or protein sequences against a target.
Small molecule screen → · Protein screen →
Next steps
Section titled “Next steps”- Get started — Create an account, install a client, and run the smallest complete example
- Authentication — API key types and SDK setup
- Agent integrations — Use Boltz from Claude Code, Codex, and other coding agents
- API Reference — Full endpoint documentation with SDK examples