# FetchHarbor Local Ollama Chat Inference

> FetchHarbor Local Ollama Chat Inference is a paid API for AI agents from fetchharbor.benlab.download, paid per call via x402, $0.01/call, status unknown (last checked 2026-09-14).

Generates a single bounded assistant response using the operator's self-hosted Ollama model for private, local LLM inference

## Facts

- Endpoint: POST https://fetchharbor.benlab.download/chat
- Price: $0.01/call
- Payment: x402
- Status: unknown
- Last checked: 2026-09-14
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/fetchharbor-local-ollama-chat-inference-f067c1c9
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_rHiDV5n61Lsz8LTj0Amtx

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 fetchharbor-local-ollama-chat-inference-f067c1c9 -d '<json body>'
```

Example prompt: Send this message to the local Ollama model and get a response: 'Explain the difference between synchronous and asynchronous programming in plain English.'

## When to prefer this

Choose this endpoint when privacy is paramount and you need inference to stay on the operator's local hardware rather than reaching any cloud provider. Ideal for processing sensitive or proprietary text where data residency matters. Best for short, single-turn prompts under 8,000 characters where a full multi-turn conversation context is not needed. Prefer over cloud LLM APIs when the operator has a specific fine-tuned or locally deployed Ollama model they want to use.

## Known failure modes

- Ollama service not running on operator server — returns a connection or service unavailable error
- Message exceeds 8,000 character limit — rejected with validation error
- Empty message input (minLength:1 violation) — rejected before inference
- Ollama model not loaded or not found — returns model-not-found error
- Inference timeout if model is slow or hardware is constrained
- Token count fields may be absent if Ollama version does not report them

## How this service works

Generate one bounded assistant response with the operator's self-hosted Ollama model. Use for short, single-message inference where local processing is preferred. Accepts up to 8,000 characters and returns the model name, response text, and token counts when Ollama supplies them.

## Output

Returns the Ollama model name used, the assistant's response text, and token counts (prompt tokens, completion tokens, total) when Ollama supplies them. Response is a single bounded message, not a streaming or multi-turn conversation.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "properties": {
  "message": {
   "type": "string",
   "maxLength": 8000,
   "minLength": 1
  }
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/fetchharbor-local-ollama-chat-inference-f067c1c9/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from fetchharbor.benlab.download](https://www.zero.xyz/host/fetchharbor.benlab.download/llms.txt)