The Neurosnap Python package supports two different kinds of work: accessing hosted computations and running scientific analysis locally. Installing the package does not automatically run hosted jobs, and local utilities do not require launching a Neurosnap service.
Choose your entry point
| Goal | Use |
|---|---|
| Submit hosted tools, monitor jobs, and retrieve outputs | The platform API and its SDK client |
| Read structures, analyze sequences, compute metrics, or work in notebooks locally | The SDK’s scientific modules |
| Connect an agent application to Neurosnap | The MCP server |
Hosted submissions use account authentication and compute credits. Local operations run in your environment and depend on the installed package and any optional scientific dependencies.
Install the Python package
Install in an isolated Python environment compatible with the SDK’s current requirements:
python -m venv .venv
source .venv/bin/activate
python -m pip install --upgrade neurosnap
The activation command above is for POSIX shells. On Windows, use the activation command for your shell. Check the SDK repository for supported Python versions, optional dependencies, and release information.
Record your installed package version with your analysis. The generated reference available on this server may describe a different version; verify names and signatures against your installation before using an example.
Automate hosted jobs
Start with the API quickstart and API reference for authentication, service discovery, multipart submissions, job monitoring, and result retrieval. The original API tutorial provides additional examples and context.
Use exact service titles and current field definitions. Never embed an API key in a shared notebook or commit it to source control. A submission response interrupted by a network failure does not prove that no job was created; inspect history before retrying.
The API client reference describes the SDK wrapper. Current endpoint behavior and service schemas should guide integrations when older examples differ.
Analyze structures and sequences locally
The generated SDK getting-started guide covers parsing structures, accessing chains and residues, rendering, alignment operations, and other scientific utilities.
Useful reference entry points:
Use the actual file and molecular representation required by an operation. A parser accepting coordinates does not establish that those coordinates have been prepared for docking or simulation.
Connect through MCP
The MCP server guide explains connection, authentication, service discovery, and job execution. The Neurosnap MCP tutorial provides additional context. The platform’s MCP endpoint is https://neurosnap.ai/mcp; follow the supported authentication flow for your client.
Inspect the available tools and their current input definitions before submitting work. Preserve returned job IDs and retrieve their status rather than submitting replacements. Review Agent data handling when deciding which clients and data are appropriate for your organization.
About the generated reference
The SDK module reference is generated from the Python package. It retains its own version information, source links, and module cross-references. Existing SDK documentation links continue to work during this documentation migration.
The published SDK reference uses the same documentation layout and search as these guides. Its package version appears above each generated article; module signatures, anchors, and source links remain generated by Sphinx. The standalone build remains available for offline use.
Platform guides are maintained separately. Reference builds are published as immutable versioned artifacts and loaded on backend startup, so rebuilding the SDK does not overwrite authored content or mix page content with assets from another release.
Try a complete workflow
The worked tutorials connect platform actions with downloadable inputs and review checkpoints. Start with sequence to structure, then use the generated scientific-module reference when adding local analysis to your own results.