> Feedback: If these docs are stale, missing, or confusing, post sanitized feedback to `https://docs.tempo.xyz/api/feedback` with `source: "mcp"`, a short `message`, and any relevant `toolName`, `relatedResource`, or `client`.
# Pay for AI models per request

Give your agents access to any LLM — OpenAI, Anthropic, Gemini, DeepSeek, Mistral, and more — without managing API keys, billing accounts, or usage limits. MPP lets agents pay per token with stablecoins on Tempo, and the model provider gets paid instantly.

## The problem

Every LLM provider requires a separate API key, billing account, and credit card on file. For a single developer this is manageable. For a fleet of autonomous agents, it's a bottleneck: each agent needs its own credentials, each provider has different billing cycles, and rate limits are tied to account tiers rather than willingness to pay.

## How MPP solves it

With MPP, your agent holds a stablecoin balance on Tempo and pays per request. No signup, no API keys, no invoices. The agent discovers the model's price via the `402` Challenge, pays with a stablecoin transfer or session voucher, and gets the response — all in a single HTTP round-trip.

**Tempo Sessions** are ideal for LLM access. The agent opens a payment channel once, then signs off-chain vouchers for each chunk of tokens received. The model provider verifies vouchers in microseconds — no blockchain calls during inference — and settles in batch later. This makes per-token billing practical without adding latency.

## Available services

| Provider | Models | Service URL |
|---|---|---|
| OpenAI | GPT-4o, o3, DALL·E, Whisper | `openai.mpp.tempo.xyz` |
| Anthropic | Claude Sonnet, Opus, Haiku | `anthropic.mpp.tempo.xyz` |
| Google Gemini | Gemini, Veo video, image gen | `gemini.mpp.tempo.xyz` |
| [DeepSeek](https://deepseek.mpp.paywithlocus.com) | DeepSeek-V3, R1 reasoning | `deepseek.mpp.paywithlocus.com` |
| [Mistral](https://mistral.mpp.paywithlocus.com) | Large, Codestral, Pixtral | `mistral.mpp.paywithlocus.com` |
| OpenRouter | 100+ models, unified API | `openrouter.mpp.tempo.xyz` |
| Grok | xAI chat, search, code exec | `grok.mpp.tempo.xyz` |
| [Perplexity](https://perplexity.mpp.paywithlocus.com) | Sonar search + grounding | `perplexity.mpp.paywithlocus.com` |

## Try it

```bash
# Install Tempo CLI + wallet
curl -L https://tempo.xyz/install | bash && tempo add request && tempo wallet login

# Call OpenAI with pay-per-request — no API key needed
tempo request openai.mpp.tempo.xyz/v1/chat/completions \
  -d '{"model": "gpt-4o", "messages": [{"role": "user", "content": "Hello"}]}'
```

## Prompt your agent

```
Use openai.mpp.tempo.xyz to call GPT-4o via Tempo.
Pay per request with stablecoins — no API key needed.
```

## Next steps

<Cards>
  <Card icon="lucide:repeat" title="Accept pay-as-you-go payments" description="Build session-based billing for your own LLM API" to="/docs/guide/machine-payments/pay-as-you-go" />

  <Card icon="lucide:book-open" title="Tempo Sessions" description="How payment channels enable per-token billing" to="https://mpp.dev/payment-methods/tempo/session" />

  <Card icon="lucide:globe" title="Browse all services" description="85+ MPP-enabled services in the directory" to="https://mpp.dev/services" />
</Cards>