Export API
The export API is intended to allow people to consume recordings and ingest them into their own storage or analytics platforms.
The main requirement is to have the same capability in the future platform, ideally backwards compatible with the old one.
We would want to add our common object model metadata on top of the backwards compatible data, that way it is easy for existing integrations to migrate and easy for new partners to write their own integrations.
Existing capabilities:
HTTP Push API, each call captured results in one POST message out of the API.
JSON style messaging with audio as an attachment.
Metadata is JSON style key value pairs
Transcription is JSON style transcript and list of words in the call
Audio is WAV currently, MP3 soon
Options:
Audio on/WAV/MP3 (MP3 is in dev, does not exist yet)
Transcriptions yes/no depending if feature is enabled
Notes:
200 response from client removes the message from internal queue, otherwise it will backoff and retry.
Questions
James Ratcliffe (Unlicensed), Simon Jolly (Unlicensed), What does Export via SMTP mean? How data should be sent via SMPT - as attachments ?
Child Pages
- Export API - Export via SMTP
- Export Engine / Transcription Integration
- Mock Transcriptor
- Speechmatics
- Call Transcription
- Export UI