Transcribe audio - /api/transcribe-audio

JWT token must contain customerId and agentId fields with purpose 'audio'. Will be sent as Authorization header.
ISO language code. If not specified, Whisper will auto-detect the language.
A text to guide the model and provide context for the transcription.


      

Extract text from any file - /api/file-to-text

Loading supported file types...
JWT token must contain customerId and agentId fields with purpose 'file'. Will be sent as Authorization header.


      

Extract text from PDF pages (image-based) - /api/pdf-page-to-text

Comma-separated page numbers to extract using vision analysis
Vision model (e.g., claude-sonnet-latest, claude-opus-latest, gpt-5.1)
Override the default prompt. E.g., "Extract only the text" or "Summarize this page"
JWT token must contain customerId and agentId fields with purpose 'file'. Will be sent as Authorization header.


      

Extract text from GCS file - /api/gcs-file-to-text

Loading supported file types...
JWT token must contain filePath with purpose 'gcs'. Will be sent as Authorization header.


      

Extract geometry from CAD files - /api/gcs-to-geojson

JWT token must contain filePath with purpose 'gcs'. Will be sent as Authorization header.


      

Extract text from URL - /api/url-to-text

Loading supported file types...
Direct URL to the file to extract text from
JWT token must contain customerId and agentId fields with purpose 'url'. Will be sent as Authorization header.