Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/UnkleFunk/HouseMusicSwarm-/llms.txt

Use this file to discover all available pages before exploring further.

The Deep Research agent is OpenSwarm’s dedicated researcher. It conducts multi-query, evidence-based research across the web and academic sources, structures findings into a consistent report format, and always backs claims with inline source citations. Every output distinguishes verified facts from speculation and presents conflicting sources side-by-side with a credibility assessment — so you get a research brief that can actually support decisions, not just a list of links.

​
What it does

Multi-query web research

Runs a minimum of 3–5 distinct search queries per request using WebSearchTool. Searches broad then narrows — official sources, regulatory filings, reputable news outlets, and industry reports take priority.

Scholar search

Uses ScholarSearch to surface peer-reviewed papers, scientific studies, and scholarly citations. Called once per request to conserve API usage.

Inline citations

Every important claim includes a source link in [Source: URL] format. No unsourced assertions are presented as facts. Long URL dumps are never included in the final output.

Balanced analysis

When sources conflict, the agent documents all perspectives, explains which appear most credible and why, and notes recency and source quality for each.

Structured report format

Every research output follows a fixed six-section structure: Executive Summary → Key Findings → Evidence and Details → Options → Recommendation → Risks and Open Questions.

Actionable options

The agent always closes with 2–4 distinct paths forward, each with pros, cons, and prerequisites — and a clear recommendation with supporting rationale.

​
Required configuration

Web search, Scholar search, and product search all require a SEARCH_API_KEY from SearchAPI.io. 
# .env
SEARCH_API_KEY=your_key_here
Get your SEARCH_API_KEY at https://www.searchapi.io. Without it the agent cannot browse the web or access Scholar search. The key enables WebSearchTool, ScholarSearch, and ProductSearch.

​
Output format

Every research response follows this structure: 
1. Executive Summary     — 5–10 actionable bullet points
2. Key Findings          — grouped by theme with clear headings
3. Evidence and Details  — inline citations, data, quotes, specific examples
4. Options               — 2–4 paths forward, each with pros/cons/prerequisites
5. Recommendation        — best option with 3–5 supporting reasons
6. Risks and Open Questions — gaps, unknowns, follow-up research suggestions

​
Example prompts

Research my top 5 competitors in the project management SaaS space. Include pricing, positioning, and recent funding.

Find recent academic papers on large language model reasoning and summarize the key findings.

What are the current trends in house music production for 2025? Focus on hardware, DAWs, and distribution channels.

Give me a comprehensive analysis of the AI agent market landscape — key players, funding rounds, and open-source alternatives.

​
Limitations

Web search, Scholar search, and product search all require a valid SEARCH_API_KEY. Without it the agent has no ability to browse the internet or access academic databases. It will clearly state this rather than fabricating results.
The agent will also explicitly state “Not found” or “Unable to verify” for any claim it cannot confirm after thorough searching — rather than speculating or filling gaps with hallucinated content. It lists exactly which searches were conducted and what information is missing. ScholarSearch is called at most once per user request to manage API costs. If your request requires extensive academic depth, consider breaking it into separate research tasks.

Build docs developers (and LLMs) love

Get started for freeTalk to us