TranscribeToolby Synapse FoundryStart with $20 free credit on first sign-in.
Then keep transcript, batch, and API work on one balance.
Transcription API
Queue async transcript jobs, request optional summary or knowledge extraction, and return transcript, HTML, Markdown, and structured knowledge artifacts from one workflow on one pricing model.
Async jobsQueue long-running transcript work without blocking your app.
Batch throughputProcess playlists, channels, interviews, and media libraries in bulk.
BillingTranscript API and batch usage are billed at $0.02/min, metered per second. Optional modules are charged only when requested, and only successfully completed transcript time deducts from balance.
ModulesRequest transcript, summary, and structured knowledge explicitly instead of paying for a forced bundle.
UseMove from source material to reviewable output faster.
AutomationFeed supported video, uploaded media, and podcast inputs into downstream research, media, and knowledge workflows.
The pricing model is the metered table below. The package cards underneath only fund prepaid balance.
Transcript$0.02/min billed per second (0.000333/sec)
Summary$0.05 per run
Knowledge extraction$0.08 per run
Batch$0.02/min billed per second across successful transcript time
These do not change the pricing contract. They only preload balance that transcript seconds and optional modules deduct from later.
Pay $25.00
$25.00 buys a total of $25.00 credits
Base transcript rate $0.02/min billed per second ($0.000333/sec)
About 20h 50m transcript time at the base rate
Pay $100.00
$10.00 bonus · $100.00 buys a total of $110.00 credits
Base transcript rate $0.02/min billed per second ($0.000333/sec)
About 91h 40m transcript time at the base rate
Pay $250.00
$35.00 bonus · $250.00 buys a total of $285.00 credits
Base transcript rate $0.02/min billed per second ($0.000333/sec)
About 237h 30m transcript time at the base rate
Pay $500.00
$100.00 bonus · $500.00 buys a total of $600.00 credits
Base transcript rate $0.02/min billed per second ($0.000333/sec)
About 500h 0m transcript time at the base rate
Because the hard part is not text generation. It is intake, batching, orchestration, packaging, and returning output you can review, search, and reuse.
ResearchTurn long-form source material into working summaries and findings.
OperationsEstimate, submit, poll, and download from one tracked flow.
ControlFund balance once, create keys, and scale workload only when needed.
Async transcript jobsYouTube, Vimeo, Bilibili, TikTok, uploaded media via media_id, and Apple or Spotify podcast episode URLs, or explicit podcast episode metadata.
Podcast contractSend one Apple or Spotify podcast episode URL in podcast_episode_url, or send explicit podcast_rss_url plus podcast_episode_url or podcast_audio_url.
Recommended routeUse /api/v1/transcript-jobs for queued end-to-end processing and poll completion.
Batch estimate + submitYouTube videos/playlists/channels, Vimeo video URLs, Bilibili video URLs, TikTok video URLs, direct podcast RSS feeds, Apple Podcasts show or episode URLs, Spotify podcast show or episode URLs, and topic_request.
Podcast noteFor batch discovery, Apple and Spotify URLs can resolve at the show or episode level.
Billing ruleOnly successfully completed transcript time deducts prepaid balance.
TrackingUse an API key to make discovery, submit, poll, and download activity attributable to a specific app or workflow.
Transcript work is metered in seconds. Optional modules are separate billable events so agents only pay for what they request.
{
"event_id": "evt_xxx",
"job_id": "job_xxx",
"account_id": "acct_xxx",
"api_key_id": "key_xxx",
"module": "transcription",
"event_type": "transcription_seconds",
"unit": "second",
"quantity": 3600,
"unit_price_microusd_numerator": 20000,
"unit_price_divisor": 60,
"amount_microusd": 1200000,
"currency": "usd",
"status": "committed",
"idempotency_key": "idem_xxx:transcription"
}Use your API key in the x-api-key header. Replace https://your-domain.com with your deployed origin.
Programmatically, treat transcript modules as explicit request fields. Batch discovery still uses /api/v1/batch-jobs/estimate, then /api/v1/batch-jobs for tracked orchestration.
curl -X POST https://your-domain.com/api/v1/batch-jobs/estimate -H "x-api-key: YOUR_API_KEY" -H "content-type: application/json" -d '{
"input_text": "https://vimeo.com/76979871"
}'TranscribeTool exposes browser-native agent tools through WebMCP. This is separate from the REST API: REST is for headless or server-to-server use, while WebMCP is for agents operating inside the signed-in website session.
DiscoveryWebMCP-capable browsers and extensions discover tools from document.modelContext on the page itself.
AuthWebsite tools use the current signed-in browser session instead of a separate website API key flow.
Current rolloutCurrent live page-bound tools include homepage batch tools get_batch_balance, start_batch_transcription, upload_media_file, and start_uploaded_media_batch_transcription; downloads tools list_batch_downloads, get_batch_job_status, and recover_batch_zip; plus transcript library/editor tools like list_transcripts, search_transcripts, open_transcript_details, get_transcript_details, list_transcript_notes, save_transcript_note, and edit_transcript.
Current scopeThese tools are page-bound progressive enhancement, not a replacement for the REST API contract.
// In a WebMCP-capable browser or extension
const tools = await document.modelContext.getTools();
// Example live tools on TranscribeTool pages
// Homepage: get_batch_balance, start_batch_transcription, upload_media_file, start_uploaded_media_batch_transcription
// Downloads: list_batch_downloads, get_batch_job_status, recover_batch_zip
// Transcript library/editor: list_transcripts, search_transcripts, open_transcript_details, get_transcript_details, list_transcript_notes, save_transcript_note, edit_transcriptThe pricing model is metered usage: transcript work is billed at $0.02/min, metered per second (~$0.000333/sec), and optional summary or knowledge runs are billed only when requested. You pay only for work that completes successfully. The funding model is prepaid balance: packages simply preload balance for those later deductions.
No. Only successfully completed transcript time deducts from balance.
No checkout happens on every call. You fund balance upfront, and only successfully completed usage deducts from that balance. Transcript seconds, summary runs, and knowledge runs are each recorded as distinct usage events.
Because it is a real product benefit. Per-second billing ensures you pay only for what you actually use, without coarse minute rounding or charges beyond successfully transcribed output.