perplexity
Perplexity AI research assistant with real-time web search, source citations, and optimized UK-centric prompts. Access live web data, current events, market research, and comprehensive analysis backed by verified sources.
$ 安裝
git clone https://github.com/pacphi/sindri /tmp/sindri && cp -r /tmp/sindri/docker/lib/extensions/vf-perplexity/resources ~/.claude/skills/sindri// tip: Run this command in your terminal to install the skill
name: perplexity description: > Perplexity AI research assistant with real-time web search, source citations, and optimized UK-centric prompts. Access live web data, current events, market research, and comprehensive analysis backed by verified sources. version: 2.0.0 author: turbo-flow-claude mcp_server: true protocol: mcp-sdk entry_point: mcp-server/server.js dependencies:
- perplexity-sdk
Perplexity AI Research Skill
Access Perplexity's powerful real-time web search and research capabilities directly from Claude Code and claude-flow.
Overview
This skill provides:
- Real-time web search via Perplexity's Sonar API
- Source citations with every response
- Prompt optimization for maximum research quality
- UK-centric focus with European context
- Multiple research modes (quick, balanced, deep)
MCP Server
The skill includes an MCP server that exposes Perplexity research tools:
Tools
-
perplexity_search - Quick factual search
- Fast, cited answers to direct questions
- Current data and live web access
-
perplexity_research - Deep research analysis
- Multi-source synthesis
- Comprehensive reports with citations
- Structured output formats
-
perplexity_generate_prompt - Prompt optimization
- Transform vague queries into effective prompts
- Apply best practices for Perplexity
- UK/European context integration
Usage from Claude Code
# Quick search
perplexity_search "current UK mortgage rates major banks"
# Deep research
perplexity_research --topic "AI trends UK enterprise 2025" \
--format "executive summary" \
--sources 10
# Generate optimized prompt
perplexity_generate_prompt --goal "market research for renewable energy ETFs" \
--context "UK retail investor £10K budget"
Configuration
API key configured via environment variable:
PERPLEXITY_API_KEY=pplx-xxxxx
Set in .env file or docker-compose environment.
Models Available
- sonar - Fast, balanced (default)
- sonar-pro - Deep research, more sources
- sonar-reasoning - Complex analysis
Features
Prompt Optimization
Automatically structures queries using five-element framework:
- Instruction - Clear goal
- Context - Background framing
- Input - Specific constraints
- Keywords - Focus terms
- Output format - Structured results
UK-Centric Research
- British English spelling/terminology
- UK laws and regulations
- European context priority
- .gov.uk and UK source preference
Citation Management
Every response includes:
- Source URLs
- Publication dates
- Credibility indicators
- Verification links
Integration with Claude Flow
Swarm Mode
cf-swarm "research UK fintech regulations using perplexity skill"
Hive Mind
cf-hive "comprehensive market analysis UK renewable energy sector" \
--tools perplexity \
--agents 5
Hooks Integration
// Pre-task hook: Generate optimized prompt
hooks.preTask((task) => {
if (task.requiresResearch) {
return perplexity.generatePrompt(task.description);
}
});
Examples
Market Research
{
"tool": "perplexity_research",
"params": {
"topic": "Top 10 UK SaaS companies by ARR 2025",
"context": "B2B software market analysis",
"format": "table",
"timeframe": "last 90 days",
"sources": 15
}
}
Regulatory Compliance
{
"tool": "perplexity_search",
"params": {
"query": "UK GDPR requirements for e-commerce 2025",
"domain_filter": [".gov.uk", ".ico.org.uk"],
"uk_focus": true
}
}
Competitive Analysis
{
"tool": "perplexity_research",
"params": {
"topic": "Compare Stripe vs GoCardless UK market share payment processing",
"format": "comparison table",
"include_pricing": true,
"uk_focus": true
}
}
Advanced Features
Search Options
- domain_filter: Limit to specific domains
- timeframe: Recency constraint (24h, 7d, 30d, 90d)
- uk_focus: Prioritize UK/EU sources
- citation_style: APA, MLA, Chicago, IEEE
- temperature: Creativity vs precision (0.0-1.0)
Output Formats
- prose: Natural language summary
- table: Structured comparison
- bullet: Concise bullet points
- executive: Executive summary with TL;DR
- report: Full research report
Rate Limiting
- 50 requests/minute (default tier)
- 500 requests/hour
- Automatic retry with backoff
- Queue management for burst requests
Troubleshooting
API Key Issues
# Test API key
curl -H "Authorization: Bearer $PERPLEXITY_API_KEY" \
https://api.perplexity.ai/chat/completions
MCP Server Not Responding
# Check supervisord status
supervisorctl status perplexity-mcp
# View logs
tail -f /var/log/perplexity-mcp.log
Rate Limit Errors
- Reduce concurrent requests
- Increase delay between calls
- Upgrade Perplexity tier
Performance
- Quick search: 2-5 seconds
- Deep research: 10-30 seconds
- Source count: 5-20 citations typical
- Memory usage: ~50MB per MCP server
Best Practices
- Be specific - Narrow scope, clear constraints
- Request citations - Always verify sources
- Use UK context - Specify regional needs
- Optimize prompts - Use generate_prompt tool first
- Check recency - Set timeframe for time-sensitive queries
- Verify sources - Review citation quality
Security
- API keys stored in environment variables only
- No credential logging
- User-specific isolation (zai-user can't access devuser keys)
- TLS for all API communication
Resources
Repository
