In molecular modeling, a task can be conceptually simple but fiddly to carry out: find the ligand, select nearby residues, change a representation, remove waters, or repeat a selection with slightly different criteria. The work is not always difficult because of the science; sometimes the friction comes from remembering which menu, editor, or command gets you there.
SAMSON AI offers an in-application way to turn plain-language requests into actions, selections, or Python scripts. The useful distinction is knowing which kind of request to use—and checking what the assistant proposes before building it into your workflow.
Start with the kind of task
Open the assistant from Interface > Assistant, or use Ctrl + 0 on Windows and Linux, or Cmd + 0 on Mac. Type / in its input field to see the available commands.
For a direct change to the open document, use /do. For example, you can ask it to zoom to the binding site, remove water, or select a ligand and apply a licorice model. This is a practical choice when you know the result you want but do not want to navigate through the interface to find each control.
For a selection defined by scientific criteria, use /select. You can express the target in natural language, such as “all lysines in chain B” or “all side chains within 5 angstrom of the ligand.” SAMSON AI displays the corresponding Node Specification Language (NSL) and executes it. Seeing the NSL gives you a chance to connect the plain-language request with the formal selection expression, which can be useful when you want to understand or refine what was selected.
When the task should become reusable
If the operation is more involved, or you may want to adapt it later, use /script to generate a Python script using the SAMSON Python API. Examples include translating all atoms along the z direction by 1 angstrom, creating a small GUI with a distance field and a Move button, or computing the receptor’s radius of gyration.
A useful rule of thumb is: use /do for an immediate document action, /select when the challenge is defining which nodes to target, and /script when you want code that you can inspect or build upon. Generated scripts are a starting point, not a substitute for checking that the operation matches your intent—especially when working with a structure you need to preserve.
Bring reference material into the conversation
Sometimes the question is not about how to operate SAMSON, but about a paper or other source you are consulting. Use /learn with a URL to have SAMSON AI read a regular web page or PDF. You can then ask questions about the learned material with /refer, including requests for a table. You may learn from multiple documents; the information remains available after SAMSON is closed until you use /forget to clear it.
This can help keep a reference and a modeling task in the same working context. Still, treat answers as assistance rather than authoritative scientific conclusions: the documentation notes that responses can be inaccurate, incomplete, or ambiguous. Verify critical details against trusted sources.
Keep privacy in mind
The SAMSON AI privacy policy says that queries and interactions may be collected and stored to improve services, that messages are encrypted during transmission, and that third-party services may be used. Avoid including sensitive information unless you are comfortable with the applicable privacy terms.
To explore the commands and their examples, read the SAMSON AI documentation.
Note: SAMSON and all SAMSON Extensions are free for non-commercial use. Get SAMSON.