API Reference

Streams

Copy Markdown

Open in Claude

Open in ChatGPT

Open in Cursor

---

Copy Markdown

View as Markdown

Add transcript

client.streams.addTranscript(stringnamespace, StreamAddTranscriptParams { stream_id, text, timestamp } body, RequestOptionsoptions?): StreamAddTranscriptResponse { requestId, status }

POST/v1/streams/{namespace}/transcripts

Ingest a single transcript segment for a livestream using a relative timestamp in milliseconds. The stream does not need to be created beforehand. Transcript timestamps are relative milliseconds from stream start and should be monotonically increasing within a stream so transcript context lines up correctly with later frames. This method returns immediately after the transcript has been accepted for processing.

ParametersExpand Collapse

namespace: string

The namespace to ingest stream data into

body: StreamAddTranscriptParams { stream_id, text, timestamp }

stream_id: string

text: string

timestamp: number

minimum0

ReturnsExpand Collapse

StreamAddTranscriptResponse { requestId, status }

Stream ingestion request accepted for asynchronous processing.

requestId: string

status: "accepted"

Add transcript

TypeScript

HTTP

TypeScript

Python

import Safetykit from 'safetykit';

const client = new Safetykit({
  apiKey: process.env['SAFETYKIT_API_KEY'], // This is the default and can be omitted
});

const response = await client.streams.addTranscript('namespace', {
  stream_id: 'YOUR_STREAM_ID',
  text: 'This is an example transcript segment.',
  timestamp: 5000,
});

console.log(response.requestId);

200 example

{
  "requestId": "req_01h2m7qdmdjckc30e1mnq6xqfd",
  "status": "accepted"
}

Returns Examples

200 example

{
  "requestId": "req_01h2m7qdmdjckc30e1mnq6xqfd",
  "status": "accepted"
}