openapi: 3.1.0
info:
  title: Kagi Search Extract API
  description: REST API for the Kagi premium search platform, exposing web search and markdown extract endpoints. All requests are authenticated with a Kagi API token issued from the Kagi API portal.
  version: 1.0.0
  termsOfService: https://kagi.com/terms
  contact:
    name: Kagi Support
    url: https://help.kagi.com/kagi/api/overview.html
  license:
    name: Proprietary
    url: https://kagi.com/terms
servers:
- url: https://kagi.com/api/v1
  description: Kagi API production endpoint
security:
- bearerAuth: []
tags:
- name: Extract
  description: Extract markdown content from web URLs.
paths:
  /extract:
    post:
      operationId: extract
      summary: Extract markdown from URLs
      description: Extract clean markdown content from up to 10 URLs per request using the Kagi extraction service.
      tags:
      - Extract
      requestBody:
        required: true
        content:
          application/json:
            schema:
              type: object
              properties:
                urls:
                  type: array
                  description: One or more URLs to extract markdown from (up to 10 per call).
                  items:
                    type: string
                    format: uri
              required:
              - urls
      responses:
        '200':
          description: Extracted markdown payload.
          content:
            application/json:
              schema:
                type: object
components:
  securitySchemes:
    bearerAuth:
      type: http
      scheme: bearer
      description: Kagi API token issued from https://kagi.com/api/keys, sent in the Authorization header as "Bearer <token>" (or the equivalent "Bot <token>" form documented for the legacy Search endpoint).