API Reference

Your API keys

Create a key below, then pass it as a Bearer token on every request. The full key is shown only once at creation — store it somewhere safe. Revoke a key any time; it stops working immediately.

Authentication

Authenticate every request with your secret key in the Authorization header. Keep your keys server-side — never ship them in a browser or mobile app.

Authorization: Bearer sky_live_xxxxxxxxxxxxxxxx

Rate limits

Requests are limited per key (default 60 per minute). Every response includes your current limit state; a 429 means you should retry after the window resets (see Retry-After).

X-RateLimit-Limit: 60
X-RateLimit-Remaining: 59
X-RateLimit-Reset: 42

Transcribe audio or video

Transcribes the media at a public URL and returns the full text plus word-level timestamps and confidence. Uses the same engine as the Skryber editor, including the automatic fallback for music-heavy audio.

Body parameters

Request

curl https://www.skryber.com/api/v1/transcribe \
  -H "Authorization: Bearer sky_live_REPLACE_WITH_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://example.com/podcast.mp3"
  }'

const res = await fetch("https://www.skryber.com/api/v1/transcribe", {
  method: "POST",
  headers: {
    "Authorization": "Bearer sky_live_REPLACE_WITH_YOUR_KEY",
    "Content-Type": "application/json",
  },
  body: JSON.stringify({ url: "https://example.com/podcast.mp3" }),
});

const data = await res.json();
console.log(data.text);

import requests

res = requests.post(
    "https://www.skryber.com/api/v1/transcribe",
    headers={"Authorization": "Bearer sky_live_REPLACE_WITH_YOUR_KEY"},
    json={"url": "https://example.com/podcast.mp3"},
)
print(res.json()["text"])

Response

{
  "id": "txn_4e1c8b2a-...",
  "object": "transcription",
  "duration": 184.32,
  "language": "en",
  "text": "Welcome back to the show. Today we're talking about...",
  "words": [
    { "word": "Welcome", "start": 0.08, "end": 0.42, "confidence": 0.99 },
    { "word": "back",    "start": 0.42, "end": 0.71, "confidence": 0.98 }
  ]
}

Viral clips pipeline

Import a video from a URL and Skryber runs the whole pipeline for you — download, transcription, AI viral-clip analysis, and auto-render of the top clips. Imports and renders are asynchronous: the call returns a job_id you poll until status is completed (queued means waiting for a worker slot — long waits are capacity, not failure).

Typical flow: POST /videos → GET /jobs/{job_id} until completed → GET /clips?video_id= → POST /clips/{id}/render for any pending clip you want as a file. The full machine-readable contract lives in the OpenAPI spec.

MCP server (for AI agents)

Every endpoint above is also exposed as a tool over the Model Context Protocol. The server is Streamable HTTP at https://www.skryber.com/api/mcp with two ways in: OAuth 2.1 — web connectors (claude.ai, ChatGPT) just add the server URL and you approve the connection in your browser — or the same Authorization: Bearer sky_live_... header as REST. Either way: identical plan, quota, and rate-limit rules. Manage OAuth-connected agents under Settings → Connected agents.

claude mcp add --transport http skryber https://www.skryber.com/api/mcp

# then run /mcp and choose Authenticate — approve it in your browser

That runs the same OAuth flow as the web connectors: you approve it once and the token is stored by your client, never written into a config file. Prefer an API key (CI, servers, anything headless)? Add the header instead — replace the placeholder with a real key, or every call comes back 401 invalid API key.

claude mcp add --transport http skryber https://www.skryber.com/api/mcp \
  --header "Authorization: Bearer sky_live_REPLACE_WITH_YOUR_KEY"

Tools: transcribe, import_video, list_videos, get_video, get_transcript, list_clips, list_caption_styles, render_clip, get_job, get_performance_metrics. Discovery card: /.well-known/mcp/server-card.json.

Errors

Errors return the matching HTTP status and a JSON body of the form { "error": { "type", "message" } }.