# three.ws LLM Proxy (x402)

> three.ws LLM Proxy (x402) is a paid API for AI agents from three.ws, paid per call via x402, $0.005/call, status unknown (last checked 2026-10-12).

Routes a text prompt to a fast or smart LLM model and returns a signed inference response with metering metadata, billed per call via x402 micropayments.

## Facts

- Endpoint: POST https://three.ws/api/x402/llm-proxy?utm_source=zero.xyz
- Price: $0.005/call
- Payment: x402
- Status: unknown
- Last checked: 2026-10-12
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/three-ws-llm-proxy-x402-a8a97afd
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_M8Jpk4Yv9CdXBtjWu8_60

Status and success rate cover calls made through Zero and Zero's own probes. Third-party monitors may report differently.

## How to call it through Zero

Zero handles the 402 payment challenge and records the run. With the Zero CLI installed (`npm i -g @zeroxyz/cli`):

```sh
zero fetch --capability three-ws-llm-proxy-x402-a8a97afd -d '<json body>'
```

Example prompt: Using the three.ws LLM proxy, send this prompt to the fast Groq model with up to 512 tokens: 'Summarize the key differences between REST and GraphQL in three bullet points.'

## When to prefer this

Choose this endpoint when you need per-call metered LLM inference billed via USDC micropayments (x402 protocol), especially when you require a cryptographically signed and auditable inference receipt. Prefer the 'fast' (Groq) option for low-latency tasks and 'smart' (Anthropic Haiku) for higher reasoning quality. Best suited for agents that need to pay only for what they use without a subscription, or that need verifiable proof of inference for compliance or logging purposes.

## Known failure modes

- Payment not provided or invalid x402 header — returns 402 Payment Required
- Prompt exceeds 4000 character limit — validation error
- max_tokens out of range (must be 1-2048) — validation error
- Model string unrecognized — may fall back to default or return error
- Groq or Anthropic upstream timeout — latency spike or 503
- Empty or missing prompt field — returns 400 Bad Request

## How this service works

Make a 3D AI agent from a selfie in 60 seconds. Drop one HTML tag to embed it on any website. Agents think on IBM watsonx.ai (Granite); optionally charge per chat with built-in micropayments.

## Output

A JSON object containing the generated text content, the model used, a unique job ID, detailed token usage (input/output/total), latency in milliseconds, the provider name, a SHA-256 hash of the prompt and response, a cryptographic response signature, and a structured inference receipt object for auditability.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "properties": {
  "model": {
   "type": "string",
   "default": "fast",
   "description": "\"fast\" (default, Groq), \"smart\" (Anthropic Haiku), or a specific model string."
  },
  "prompt": {
   "type": "string",
   "maxLength": 4000,
   "minLength": 1,
   "description": "The user prompt to complete."
  },
  "max_tokens": {
   "type": "integer",
   "default": 256,
   "maximum": 2048,
   "minimum": 1,
   "description": "Maximum tokens to generate."
  }
 }
}
```

## Response schema (JSON Schema)

```json
{
 "type": "json",
 "example": {
  "model": "llama-3.3-70b-versatile",
  "job_id": "9b2f1c4e-…",
  "content": "1, 2, 3.",
  "metering": {
   "type": "three-inference-response/v1",
   "jobId": "9b2f1c4e-…",
   "model": "llama-3.3-70b-versatile",
   "route": "/api/x402/llm-proxy",
   "provider": "groq",
   "latencyMs": 312,
   "tokensUsed": 11,
   "inputTokens": 5,
   "outputTokens": 6,
   "promptSha256": "a1b2…",
   "responseSha256": "c3d4…"
  },
  "provider": "groq",
  "latency_ms": 312,
  "tokens_used": 11,
  "input_tokens": 5,
  "output_tokens": 6,
  "response_signer": "NodePub…",
  "inferenceReceipt": {
   "receiptType": "three-inference-receipt/v1"
  },
  "response_signature": "4xY…"
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/three-ws-llm-proxy-x402-a8a97afd/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from three.ws](https://www.zero.xyz/host/three.ws/llms.txt)