Research Agent - MINDBLIX Documentation

Overview

The Research Agent is designed for deep information gathering and synthesis. It can search the web, analyze documents, and produce comprehensive research reports with cited sources.

Property	Value
Template ID	`research`
Default Model	Gemini 1.5 Flash
Avg. Response Time	~8 seconds
Accuracy Rate	94%

Capabilities

🔍 Web Search — Search the web for relevant information using multiple search engines
📄 Document Analysis — Extract and analyze content from PDFs, DOCX, and web pages
✅ Fact-Checking — Cross-reference claims against multiple sources
📊 Competitive Intelligence — Analyze competitors, market trends, and industry reports
📝 Report Generation — Produce structured research reports with citations

Usage

API
Python

curl -X POST http://localhost:8000/api/v1/agents/research/execute \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{
    "query": "What are the latest trends in AI agent orchestration?",
    "config": {
      "max_sources": 10,
      "output_format": "report"
    }
  }'

from mindblix import MindBlix

client = MindBlix(api_key="YOUR_API_KEY")

result = client.agents.execute(
    template="research",
    query="What are the latest trends in AI agent orchestration?",
    config={"max_sources": 10, "output_format": "report"}
)

print(result.report)

Output Example

{
  "summary": "AI agent orchestration is rapidly evolving with...",
  "key_findings": [
    {
      "finding": "Multi-agent systems are replacing single LLM calls",
      "confidence": 0.94,
      "sources": ["arxiv.org/...", "blog.google/..."]
    }
  ],
  "sources": [
    {
      "url": "https://arxiv.org/abs/2024.xxxxx",
      "title": "Survey of Multi-Agent AI Systems",
      "relevance_score": 0.95
    }
  ],
  "metadata": {
    "model": "gemini-1.5-flash",
    "tokens_used": 4521,
    "execution_time_ms": 7840
  }
}