Pricing

Pay per minute.
That's it.

No subscriptions, no tiers, no surprise bills. Every endpoint has a fixed per-minute rate. Your agent sees the price before it pays.

Inspect
$0.001 / req

Per-request pricing. Analyze quality before processing.

  • — File metadata & format
  • — Noise, clipping, silence
  • — Transcription readiness
  • — Preflight & pipeline rec.
Transform
$0.003 / min

Per-minute pricing. Each operation is a separate endpoint.

  • — Denoise ($0.01/min)
  • — Normalize loudness
  • — Trim edges & silences
  • — Change speed, transcode
Structure
$0.01 / min

Segment and label audio by speaker and speech activity.

  • — Voice activity detection
  • — Speaker diarization
  • — Speech/silence ratio
  • — Segment timestamps
Transcribe
$0.05 / min

Multi-provider STT with automatic fallback.

  • — OpenAI Whisper
  • — ElevenLabs Scribe v2
  • — Gemini Flash
  • — Multi-language, async

Always Free

Free endpoints

GET /healthz

Service status. Always returns 200 when operational.

GET /v1/catalog

All available operations with pricing. Machine-readable.

POST /v1/quotes

Get a cost estimate for any operation before processing.

POST /v1/assets

Upload, manage, and download audio assets.

GET /v1/jobs/:id

Poll async job status, cancel, or get receipt.

GET /v1/openapi.json

Full OpenAPI 3.1 spec for agent auto-discovery.

How Payment Works

x402 protocol

SoundHalo uses the x402 payment protocol — an extension of HTTP that adds a native payment layer. No API keys. No accounts. No billing dashboards.

When your agent hits a paid endpoint, it receives a 402 Payment Required response containing the exact cost. The agent signs a stablecoin payment and re-sends the request. Processing begins immediately.

Think of it like HTTP auth, but with money. Purpose-built for autonomous agent commerce where there's no human to manage credentials.

x402 payment flow 
# 1. Agent sends request
POST /v1/extract/transcribe
→ 402 Payment Required
→ X-Payment-Amount: 1.54 USDC

# 2. Agent signs and pays
POST /v1/extract/transcribe
X-Payment-Proof: <signed-tx>

# 3. Processing begins
→ 200 OK
→ { "text": "Welcome to...", "segments": [...] }

Pricing FAQ

How is audio duration measured?

We measure the actual duration of the audio content in your file, rounded up to the nearest second. Silence at the beginning and end is included in the duration.

Can I see the cost before paying?

Always. Use the free /v1/quotes endpoint to get an exact cost estimate for any operation. The x402 protocol also returns a price quote with every 402 response before any payment.

What currency do you accept?

All payments are in USDC stablecoin via the x402 protocol. Your agent needs a funded wallet — no credit card or invoicing.

Are there volume discounts?

Not yet. Pricing is flat and transparent. We may introduce volume tiers in the future, but the per-minute rate will always be available.

What if processing fails?

You're not charged. Payment is only captured after successful processing. If the service errors, the payment is not settled.