Skip to content

API Reference

Uspeech Analytics exposes a REST API so you can drive the platform from your own scripts, services, and integrations. Everything you can do in the web app — upload audio, transcribe, list and download transcripts — has an API counterpart.

The reference covers the Transcription surface end-to-end, plus the call center surface: attaching metadata to recordings and reading the analyzed results back, and the Projects endpoints you need to look up a project ID. More endpoint groups (analysis configuration, survey processing) will be added here as they stabilize.


All endpoints are mounted under /api/ on your Uspeech host:

https://app.uspeech.io/api/

If you self-host, replace the host with your own deployment URL.


Every endpoint requires either a logged-in browser session or a user-scoped API key. For programmatic access, use an API key — see Authentication.


  • Authentication — How to create API keys, the header format, and how keys are scoped to a user’s projects.
  • Projects — List the projects a key can see and find the project ID that every other endpoint needs.
  • Transcription — Upload audio, poll status, list and retrieve transcripts, and download SRT or DOCX output.
  • Uploading Calls for Analysis — Upload recordings into a project with automatic analysis enabled, track batch progress, and read results back.
  • Call Metadata — Attach, update and read back call date/time, agent, direction and campaign — per file or as a CSV manifest.
  • Call Reports — Read call center results back: aggregated report metrics, the per-call explorer, analysis progress, and the PDF export.
  • Status Codes & Errors — HTTP status codes you should expect, the shape of error responses, and transcription-specific failure reasons.

A live, auto-generated OpenAPI 3 schema is available for every deployment:

  • OpenAPI documentGET /api/schema/
  • Swagger UI/api/schema/swagger-ui/
  • ReDoc/api/schema/redoc/

These pages reflect the exact endpoints, parameters, and response shapes in your installed version. Use this written reference for the “how do I do X end-to-end” walkthrough, and the schema for the precise list of fields on any individual endpoint.


  • Request and response bodies are JSON unless stated otherwise. File uploads use multipart/form-data.
  • Endpoint headers are written as ## METHOD /path.
  • Every example is available as cURL, Python and JavaScript — pick a language with the tabs above any example. Your choice is remembered and applies to every example on every page.
  • Examples read the key from a USPEECH_KEY environment variable and send it as Authorization: Api-Key …. Substitute your own host and key.
  • The Python examples use requests; the JavaScript examples use the built-in fetch and run on Node 18 or newer.

💡 Tip: every example here assumes you’ve already created an API key — start at Authentication if you haven’t.