playbooks

home / skills / openclaw / skills / parallel

parallel skill

/skills/mvanhorn/parallel

This skill enables high-accuracy web research for AI agents using the Parallel.ai API, returning rich excerpts and citations from private and public sources.

npx playbooks add skill openclaw/skills --skill parallel

Review the files below or copy the command above to add this skill to your agents.

Files (9)
SKILL.md
2.5 KB
---
name: parallel
version: "1.1.0"
description: High-accuracy web search and research via Parallel.ai API. Optimized for AI agents with rich excerpts and citations. Supports agentic mode for token-efficient multi-step reasoning.
author: mvanhorn
license: MIT
repository: https://github.com/mvanhorn/clawdbot-skill-parallel
homepage: https://parallel.ai
triggers:
  - parallel
  - deep search
  - research
metadata:
  openclaw:
    emoji: "🔬"
    requires:
      env:
        - PARALLEL_API_KEY
    primaryEnv: PARALLEL_API_KEY
    tags:
      - search
      - research
      - web
      - parallel
      - citations
---

# Parallel.ai 🔬

High-accuracy web search API built for AI agents. Outperforms Perplexity/Exa on research benchmarks.

## Setup

```bash
pip install parallel-web
```

API key is configured. Uses Python SDK.

```python
from parallel import Parallel
client = Parallel(api_key="YOUR_KEY")
response = client.beta.search(
    mode="one-shot",  # or "fast" for lower latency/cost, "agentic" for multi-hop
    max_results=10,
    objective="your query"
)
```

## Modes

| Mode | Use Case | Tradeoff |
|------|----------|----------|
| `one-shot` | Default, balanced accuracy | Best for most queries |
| `fast` âš¡ | Quick lookups, cost-sensitive | Lower latency/cost, may sacrifice some accuracy |
| `agentic` | Complex multi-hop research | Higher accuracy, more expensive |

## Quick Usage

```bash
# Default search (one-shot mode)
{baseDir}/.venv/bin/python {baseDir}/scripts/search.py "Who is the CEO of Anthropic?" --max-results 5

# Fast mode - lower latency/cost âš¡
{baseDir}/.venv/bin/python {baseDir}/scripts/search.py "latest AI news" --mode fast

# Agentic mode - complex research
{baseDir}/.venv/bin/python {baseDir}/scripts/search.py "compare transformer architectures" --mode agentic

# JSON output
{baseDir}/.venv/bin/python {baseDir}/scripts/search.py "latest AI news" --json
```

## Response Format

Returns structured results with:
- `search_id` - unique search identifier
- `results[]` - array of results with:
  - `url` - source URL
  - `title` - page title
  - `excerpts[]` - relevant text excerpts
  - `publish_date` - when available
- `usage` - API usage stats

## When to Use

- **Deep research** requiring cross-referenced facts
- **Company/person research** with citations
- **Fact-checking** with evidence-based outputs
- **Complex queries** that need multi-hop reasoning
- Higher accuracy than traditional search for research tasks

## API Reference

Docs: https://docs.parallel.ai
Platform: https://platform.parallel.ai

Overview

This skill provides high-accuracy web search and research powered by the Parallel.ai API, optimized for AI agents that need rich excerpts and citations. It supports quick lookups, deep multi-hop research, entity discovery, clean content extraction from URLs/PDFs, and continuous monitoring. Optional authenticated-source access enables research across private or paywalled resources.

How this skill works

The skill wraps Parallel.ai endpoints (Search, Task, FindAll, Extract, Monitor) to perform targeted research workflows. Queries are routed to processors tuned for speed or depth (base, core, ultra) and results come with structured excerpts and citations. For authenticated/private sources, the skill uses browser-use.com MCP integration when BROWSERUSE_API_KEY is provided.

When to use it

  • Quick fact checks or current-events lookups where speed matters (Search).
  • Deep, cited research reports or structured enrichment for companies and documents (Task).
  • Generating validated lists of entities (leads, companies, practitioners) at scale (FindAll).
  • Converting web pages or PDFs into clean markdown and focused excerpts (Extract).
  • Continuous tracking and alerts for news, prices, or policy changes (Monitor).

Best practices

  • Choose the appropriate API: Search for speed, Task for depth, FindAll for discovery, Extract for documents, Monitor for recurring tracking.
  • Use processor selection (base/core/ultra) to balance cost and depth; start with core for most use cases.
  • Provide clear objectives for Extract (e.g., 'Extract risk factors') to improve focus and output quality.
  • Limit FindAll jobs with enrich/limit flags to control cost and runtime; use pro generator only when maximum recall is required.
  • Store and review citations from outputs for compliance and auditability when using results in reports or downstream systems.

Example use cases

  • Generate a cited market analysis report on HVAC industry trends in the USA (Task --report).
  • Build a prospect list of residential roofing companies in Charlotte, NC with contact and review data (FindAll + enrich).
  • Extract API docs or full PDFs into clean markdown for developer onboarding (Extract --full).
  • Set a daily monitor for competitor product announcements or price drops with webhook alerts (Monitor).
  • Enrich company profiles (founding year, funding, headcount) for CRM import (Task --enrich).

FAQ

Do I need an API key?

Yes. Set PARALLEL_API_KEY for all basic functionality. BROWSERUSE_API_KEY is optional for authenticated sources.

Can this access paywalled or internal sites?

Yes, when you configure browser-use.com MCP and provide BROWSERUSE_API_KEY, authenticated source access is supported.