Skip to main content

​
Base URL

https://skilldiscs.com/api/v1

​
MCP Endpoint

https://skilldiscs.com/api/mcp

​
Endpoints

MethodPathPurpose
POST/searchSemantic search across owned Disks
GET/disksList & filter owned and active-saved Disks
GET/disks/{id}Read one Disk (UUID or public_id)
POST/disks/batchRead up to 10 owned Disks in one call
POST/api/mcpMCP Streamable HTTP endpoint for read-only knowledge tools

​
Identifiers

Disks have two identifiers:
  • id — internal UUID (e.g. f47ac10b-58cc-4372-a567-0e02b2c3d479).
  • public_id — short URL-safe slug (e.g. abc123).
Endpoints accepting {id} accept either form.

​
Response shape

All responses are JSON. Errors follow this shape: 
{
  "error": "rate_limit_exceeded",
  "message": "Optional human description",
  "retry_after": 42
}

​
Redaction policy

The API never assumes all visible Disks can return verbatim source text.
  • Owner reading own private Disk: full text and sections[].text.
  • Owner reading own published Disk: redacted: true; sections include key points but not full text.
  • Active-saved Disk: redacted: true; metadata and AI-derived fields only.
Search hits use the same policy: a hit contains either section_text or snippet, never both.

​
CORS

CORS is enabled. Browser clients on any origin can call the API directly — but never embed your API key in a client-side bundle. Proxy through your own backend.