# Parallel Web Systems Async AI Task

> Parallel Web Systems Async AI Task is a paid API for AI agents from Parallel at parallelmpp.dev, paid per call via MPP or x402, $0.3/call, status unknown (last checked 2026-10-11).

Starts an asynchronous AI-powered web research or extraction task, returning a run ID to poll for results

## Facts

- Endpoint: POST https://parallelmpp.dev/api/task?utm_source=zero.xyz
- Price: $0.3/call
- Payment: MPP, x402
- Status: unknown
- Last checked: 2026-10-11
- Activations on Zero: 1
- Provider: Parallel
- Docs: https://parallelmpp.dev/#agents
- Website: https://parallel.ai
- Tags: answers, search, web, extraction, research, ai, x402
- Canonical page: https://www.zero.xyz/c/parallel-task-3e7e7b47
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_T7A8_NyWsxZXgqAgIawDN

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 parallel-task-3e7e7b47 -d '<json body>'
```

Example prompt: Use Parallel's ultra tier to research the latest funding rounds and key investors for AI infrastructure startups in 2024 — give me a comprehensive summary when it's done.

## When to prefer this

Choose this endpoint when you need comprehensive, multi-hop AI web research or extraction that may take more than a few seconds and is better handled asynchronously. Use 'ultra' ($0.30) for highest-quality, deep research tasks and 'pro' ($0.10) for standard quality. Prefer this over the synchronous Search or Extract endpoints when the task is complex, requires agent-like reasoning across multiple web sources, or when you can tolerate polling for results rather than needing an immediate response.

## Known failure modes

- Missing required 'processor' field returns 400 validation error
- Invalid processor tier value (not 'ultra' or 'pro') returns 400
- Neither 'input' nor 'query' provided may result in empty or error response
- Insufficient funds / payment failure returns 402
- Task run times out or fails internally — poll response will reflect error state
- Rate limiting may return 429 if too many concurrent tasks

## How this service works

Run an AI research task

## Output

A run_id string that can be polled via GET /api/task/{runId} until the async task completes; final result contains AI-synthesized web research or extracted content based on the query and selected processor tier (ultra or pro).

## Example request

```json
{
 "query": "artificial intelligence research trends",
 "processor": "pro"
}
```

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "required": [
  "processor"
 ],
 "properties": {
  "input": {
   "type": "string",
   "description": "Task input (or use query)"
  },
  "query": {
   "type": "string",
   "description": "Task query (shorthand for input)"
  },
  "processor": {
   "enum": [
    "ultra",
    "pro"
   ],
   "type": "string",
   "description": "Task tier: ultra ($0.30) or pro ($0.10)"
  }
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/parallel-task-3e7e7b47/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from parallelmpp.dev](https://www.zero.xyz/host/parallelmpp.dev/llms.txt)