Parallel Deep Research

Parallel Deep Research

Comprehensive intelligence reports from natural language queries using Parallel AI Task API.

When to Use

• User needs thorough, multi-source research on a topic
• User asks complex questions requiring synthesis of multiple sources
• User needs market research or competitive analysis
• User wants comprehensive reports with citations
• Simple web search is insufficient for the depth needed

Requirements

Environment variable must be set:

• PARALLEL_API_KEY - Parallel AI API key

Command

./scripts/parallel_research research "query" [options]

Options

Processors

Output Format

Default JSON for LLM parsing:

{
  "query": "What are the latest developments in fusion energy?",
  "processor": "pro-fast",
  "run_id": "run_abc123",
  "content": {
    "summary": "...",
    "key_developments": [...],
    "companies": [...],
    "timeline": [...]
  },
  "basis": [
    {
      "field": "summary",
      "excerpts": [
        {"url": "https://...", "text": "..."}
      ]
    }
  ]
}

Examples

Basic research:

./scripts/parallel_research research "What are the key players in the AI chip market and their competitive positions?"

Higher quality (slower):

./scripts/parallel_research research "Comprehensive analysis of renewable energy trends 2024" -p ultra

Human-readable output:

./scripts/parallel_research research "Electric vehicle battery technology advances" -f markdown

Extended timeout for complex research:

./scripts/parallel_research research "Deep competitive analysis of cloud providers" -t 1800

Notes

• Research can take 1-45 minutes depending on complexity
• The pro model and ultra models are expensive and can take 10-45 minutes to complete
• Query must be under 15,000 characters
• Status updates print to stderr, results to stdout
• basis field contains citations mapping content to sources

Source: nc9/skills — distributed by TomeVault.