# TokenGuard LLM Chat Completion

> TokenGuard LLM Chat Completion is a paid API for AI agents from eltociear-tokenguard.hf.space, paid per call via x402, $0.05/call, status unknown (last checked 2026-09-15).

Runs a chat completion against an open-weight 20B-class language model with automatic failover across multiple inference backends

## Facts

- Endpoint: POST https://eltociear-tokenguard.hf.space/llm/chat
- Price: $0.05/call
- Payment: x402
- Status: unknown
- Last checked: 2026-09-15
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/tokenguard-llm-chat-completion-3a2eb56a
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_rLDiOC8TMNBTRCPoL95xE

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 tokenguard-llm-chat-completion-3a2eb56a -d '<json body>'
```

Example prompt: Using the open-weight 20B model with failover, answer this prompt: 'Summarize the key risks of DeFi yield farming in 3 bullet points' — keep the response under 300 tokens and use a temperature of 0.5.

## When to prefer this

Choose this endpoint when you need a reliable open-weight 20B-class model that won't go down if a single provider is rate-limited or unavailable, and when you want a simple prompt-in/text-out interface without managing multiple API keys. Prefer it over single-backend LLM APIs for agent workflows that require high availability, or when you want an open-weight alternative to proprietary models like GPT-4 or Claude for cost-sensitive or privacy-conscious use cases.

## Known failure modes

- All inference backends simultaneously unavailable or overloaded — returns an error with no generated text
- Prompt exceeds model context window — may truncate or return an error
- Invalid temperature or max_tokens values outside accepted ranges — likely a 400 validation error
- Payment failure via x402 protocol — call rejected before reaching the model
- Empty or malformed prompt field — model may return degenerate or empty output

## How this service works

Run a chat completion against an open-weight 20B-class model — plain prompt in, text out, with automatic failover across independent inference backends so a single call still answers when any one provider is down or rate-limiting

## Output

A plain text string containing the model's generated response to the provided prompt, shaped by the optional system instruction, max_tokens limit (1–4000, default 800), and temperature setting (0–2, default 0.3). The response is returned from whichever available inference backend successfully processed the request after automatic failover.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "properties": {
  "prompt": {
   "type": "string",
   "description": "The user prompt"
  },
  "system": {
   "type": "string",
   "description": "Optional system instruction"
  },
  "max_tokens": {
   "type": "integer",
   "description": "1-4000 (default 800)"
  },
  "temperature": {
   "type": "number",
   "description": "0-2 (default 0.3)"
  }
 }
}
```

## Response schema (JSON Schema)

```json
{
 "type": "json",
 "example": {
  "usage": {
   "ms": 740,
   "total_tokens": 210
  },
  "answer": "EIP-3009 lets a holder sign an off-chain authorization…",
  "provider": "llm7"
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/tokenguard-llm-chat-completion-3a2eb56a/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from eltociear-tokenguard.hf.space](https://www.zero.xyz/host/eltociear-tokenguard.hf.space/llms.txt)