# ds402 — DeepSeek Flash pay-per-call via x402

> ds402 — DeepSeek Flash pay-per-call via x402 is a paid API for AI agents from ds.xhclco.com, paid per call via x402, $0.01/call, status unknown (last checked 2026-11-04).

Serves DeepSeek Flash chat completions on a pay-per-call basis using x402 micropayments — no API keys or accounts required, just $0.01 USDC per request.

## Facts

- Endpoint: POST https://ds.xhclco.com/v1/flash/completions?utm_source=zero.xyz
- Price: $0.01/call
- Payment: x402
- Status: unknown
- Last checked: 2026-11-04
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/ds402-deepseek-flash-pay-per-call-via-x402-9ec2d4b7
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_x2o1Ojp8kJgdf-3oHi0zs

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 ds402-deepseek-flash-pay-per-call-via-x402-9ec2d4b7 -d '<json body>'
```

Example prompt: Send the message 'Summarize the pros and cons of nuclear energy in 3 bullet points' to the DeepSeek Flash model via the pay-per-call x402 endpoint — no API key, just pay $0.01 USDC on Base — and give me back the response, using a max of 512 tokens.

## When to prefer this

Prefer this endpoint when you need a one-off or low-volume DeepSeek Flash inference call and want to avoid account creation, API key management, or monthly subscriptions. Ideal for crypto-native agents, serverless pipelines, or agentic workflows where per-call USDC micropayments on Base are acceptable. Not suitable for high-volume inference (cost adds up vs. subscription), streaming use cases, or scenarios requiring models other than DeepSeek Flash.

## Known failure modes

- HTTP 402 returned if x402 payment challenge not satisfied — agent must detect and handle the payment flow before retrying
- max_tokens clamped to [128, 2048] — requests outside this range are silently adjusted, potentially truncating long outputs
- Streaming (stream: true) is not supported and will fail or be ignored — must use non-streaming mode
- Empty or truncated content if max_tokens is set below 256, since the model performs reasoning before outputting text
- Payment settlement failure on Base network (e.g. insufficient USDC balance, wrong chain) results in no completion being returned
- Model parameter is ignored — always uses deepseek-flash, so callers expecting a different model will silently get flash

## How this service works

Pay-per-request DeepSeek inference behind an x402 payment wall. No accounts, no API keys: POST OpenAI-style chat completions, receive HTTP 402 with an x402 v2 challenge, settle $0.01 (flash) or $0.04 (v4-pro) in USDC on Base, get the completion. One payment per call.

## Output

Returns an OpenAI-compatible chat completion object containing the assistant's message content, finish reason, and token usage counts. Streaming is not supported; the full response is returned synchronously after the x402 payment is settled.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "required": [
  "messages"
 ],
 "properties": {
  "model": {
   "type": "string",
   "description": "ignored, pinned to deepseek-flash by route"
  },
  "stream": {
   "type": "boolean",
   "description": "must be false; SSE streaming unsupported"
  },
  "messages": {
   "type": "array"
  },
  "max_tokens": {
   "type": "integer",
   "description": "clamped to [128,2048]; model does reasoning, use >=256 to avoid empty content"
  },
  "temperature": {
   "type": "number"
  }
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/ds402-deepseek-flash-pay-per-call-via-x402-9ec2d4b7/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from ds.xhclco.com](https://www.zero.xyz/host/ds.xhclco.com/llms.txt)