Cloud Speech

Service overview

What is Cloud Speech?

Google Cloud Speech-to-Text and Text-to-Speech in one bundle. 125+ languages, real-time streaming, neural voices. Robots transcribe customer orders, give voice prompts, or run wake-word loops.

Endpoints

• POST/v1/speech:recognize

  Transcribe audio (sync, up to 60s).

  $0.004 / sec
• POST/v1/text:synthesize

  Generate speech audio from text.

  $0.0001 / char

FAQ

Questions agents ask about Cloud Speech

• Does it require an API key?+

  No. Robots pay per request via x402 or pay.sh — no signup, no key, no rate limits beyond solvency.

• How does payment work?+

  Each request is signed and paid in USDC from your robot’s wallet at call time. The catalog gateway settles and forwards.

• How deterministic are responses?+

  Depends on the upstream service. LLM endpoints expose temperature; deterministic services return the same payload on the same input.

• What happens if payment fails?+

  The endpoint returns 402 Payment Required and the agent retries with adjusted funding or a different service.

• What if dynamic pricing changes?+

  The 402 response carries the current price. Robots can accept, reject, or fall back to a cheaper substitute in the same category.