# DexL Agents AI Gateway

> DexL Agents AI Gateway is a paid API for AI agents from agents.dexl.io, paid per call via x402, $0.001487/call, status unknown (last checked 2026-09-14).

Routes chat completion requests to Google Gemini models via a single pay-per-call API gateway using x402 micropayments — no accounts or API keys required.

## Facts

- Endpoint: POST https://agents.dexl.io/v1/chat/completions
- Price: $0.001487/call
- Payment: x402
- Status: unknown
- Last checked: 2026-09-14
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/dexl-agents-ai-gateway-209e970d
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_hUpfySsW749N9ZpGOXCGC

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 dexl-agents-ai-gateway-209e970d -d '<json body>'
```

Example prompt: Using DexL's pay-per-call gateway with gemini-3.5-flash, generate a 200-token summary of the following text: 'Artificial intelligence is transforming industries by automating complex tasks that previously required human expertise.'

## When to prefer this

Choose this endpoint when you need to access Gemini models without managing API keys or accounts, especially in machine-to-machine or agent workflows where x402 micropayment-per-call billing is preferred over subscription plans. Ideal for serverless agents, one-off inference calls, or multi-tenant systems where per-request cost attribution matters.

## Known failure modes

- Payment not received or x402 payment rejected — request blocked before inference
- Requested model name not in enum (gemini-3.6-flash, gemini-3.7-flash, gemini-3.5-flash, gemini-3.5-flash-lite) — validation error
- max_tokens not specified or set too high — quote computation fails or request rejected
- Upstream Gemini model unavailable or rate-limited — gateway returns 5xx error
- Malformed messages array — request rejected with schema validation error

## How this service works

Chat completions for GPT-5.6, DeepSeek V4 and Gemini behind one OpenAI-shaped endpoint. No account, no API key: the request pays for itself in USDC on Base, Polygon, Arbitrum or Solana. Pay with the upto scheme and only what you actually use settles - the unused ceiling comes back.

## Output

A chat completion response following the OpenAI-compatible format, containing the assistant's generated text message, model used, token usage counts (prompt, completion, total), and finish reason.

## Response schema (JSON Schema)

```json
{
 "type": "json",
 "example": {
  "id": "chatcmpl-...",
  "usage": {
   "total_tokens": 10,
   "prompt_tokens": 8,
   "completion_tokens": 2
  },
  "choices": [
   {
    "message": {
     "role": "assistant",
     "content": "Hello."
    }
   }
  ]
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/dexl-agents-ai-gateway-209e970d/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from agents.dexl.io](https://www.zero.xyz/host/agents.dexl.io/llms.txt)