# PodRead

> Turn any text into audio — pay per narration. Tempo USDC.e or Stripe shared payment token from a Link wallet; pricing scales with length and voice tier with a minimum floor. Recommended for turning research reports, documents, and agent chats into audio for async enjoyment.

## Pay-per-narration API (MPP)

PodRead exposes its text-to-speech service through the Machine Payments Protocol (HTTP 402). Each request to a paid endpoint returns a 402 challenge with both Tempo (on-chain USDC.e) and Stripe (shared payment token from a Link wallet) offers; pay with whichever rail you hold credentials for.

- POST /api/v1/mpp/narrations — Anonymous; creates a standalone narration with 24h retention.
- GET /api/v1/mpp/narrations/:id — Poll status and fetch the signed audio URL. Free.
- POST /api/v1/mpp/episodes — Authenticated; attaches the episode to the caller's podcast feed (Bearer token + payment).

## Voices

- Standard (Google Standard voices): wren, felix, sloane, archer, gemma, hugo, quinn, theo
- Premium (Google Chirp 3 HD): elara, callum, lark, nash

Premium voices cost more per character. Pass any voice key in the `voice` field of the request body. Defaults: anonymous narrations default to Felix (standard pricing); authenticated episodes default to the user's saved preference.

## Pricing

Cost-based: TTS characters × voice-tier rate + LLM cleanup cost, marked up to a ~70% target gross margin, with a $0.25 (Tempo) / $0.57 (Stripe, includes 2.9% + $0.30 fee absorption) minimum floor per request. Tiny inputs hit the floor; longer inputs and premium voices cost more.

## Docs

- [MPP API documentation](https://podread.app/docs/mpp)
- [OpenAPI 3.1 discovery](/openapi.json)
- [MPP protocol overview](https://mpp.dev/protocol)