Get started
1
Get sandbox keys
Create a token from Get API keys.
Sandbox data is synthetic, so no PHI leaves your machine while you build.
2
Point your agent at the docs
Add the agent instructions to a Cursor rule,
Claude skill, or the first message of a chat.
3
Paste a flow prompt
Start with First referral, end to end, then
add webhooks and status tracking.
AI-friendly resources
OpenAPI contract
Every endpoint, parameter, and schema. Use it when the agent generates
requests, clients, or validators.
llms.txt
An index of every page with a one-line description. Many agents look for it
on their own.
llms-full.txt
The full text of every page in one file. Use it when the agent needs the
rules behind the endpoints, such as status, webhooks, and errors.
Markdown per page
Add
.md to any page URL. Use it when a token limit rules out the full
file.Copy page menu
The menu at the top of every page copies it as markdown or opens it in
ChatGPT, Claude, Perplexity, Cursor, or VS Code.
MCP server
https://docs.scan.com/mcp lets an MCP client search these docs. Connect
it from the Copy page menu.Add it to your agent
Save this as a Cursor rule or Claude skill so every session starts with it.Flow prompts
Each prompt names the pages to read, the calls in order, and the rules agents most often get wrong. Replace the language in the first line with your own.First referral, end to end
Webhook receiver
Status tracking
Reports, DICOM, and invoices
Structured errors
Lines of business
Add one of these after the first-referral prompt.Example prompts
Once the agent has the docs, short asks work:- “Generate a Ruby client for creating a referral with retries on 5xx.”
- “Which fields on create are required for a workers’ comp referral?”
- “Write a TypeScript type for the visit.status_changed webhook payload from the OpenAPI spec.”
- “Why would POST /referrals return validation_error, and how do I fix it?”

