Perplexity

AI-powered search and research with real-time web access.

Provider: Perplexity AI
Authentication: API Key required
Category: AI & Research
Credit Cost: 3 credits per query

Overview

Perplexity tools provide AI-powered search capabilities that combine real-time web search with large language model reasoning. Get answers grounded in current sources with citations, making it ideal for research, fact-checking, and knowledge synthesis.

Key Features

Real-time Web Access: Searches the internet for current information
Source Citations: Answers include links to source materials
Specialized Modes: Academic papers, arXiv, SEC filings, Wikipedia
Domain Filtering: Restrict searches to specific domains
Date Filtering: Focus on content from specific time periods

Setup

Get Perplexity API Key

Go to Perplexity API Settings
Click "Generate API Key"
Copy the generated key (starts with pplx-)

Add to Reeva

Dashboard → Accounts → Add Account
Select Perplexity
Paste your API key
Click Save

Available Tools

Perplexity Ask

General-purpose AI question answering with web search.

Tool ID: Perplexity_Ask
Credit Cost: 3 credits

Parameters:

question (string, required): What you want Perplexity to answer
- Length: 5-2000 characters
model (string, optional): Model to use
- Default: "sonar-pro"
max_tokens (integer, optional): Maximum tokens to generate
- Default: 1024
- Range: 1-8192
temperature (number, optional): Sampling temperature
- Default: 0.2
- Range: 0-1 (lower = more focused, higher = more creative)
system_prompt (string, optional): Custom system guidance
- Maximum: 4000 characters

Response:

{
  "answer": "The James Webb Space Telescope (JWST) is a space telescope...",
  "citations": [
    {
      "url": "https://webb.nasa.gov/",
      "title": "James Webb Space Telescope - NASA"
    },
    {
      "url": "https://www.space.com/jwst-discoveries",
      "title": "JWST Discoveries - Space.com"
    }
  ],
  "model": "sonar-pro",
  "usage": {
    "prompt_tokens": 45,
    "completion_tokens": 256
  }
}

Example Usage:

# Python - Basic question
response = client.call_tool(
    name="Perplexity_Ask",
    arguments={
        "question": "What are the latest developments in quantum computing?"
    }
)

print(response["answer"])
print("\nSources:")
for citation in response.get("citations", []):
    print(f"  - {citation['title']}: {citation['url']}")

// TypeScript - With custom system prompt
const response = await client.callTool({
  name: "Perplexity_Ask",
  arguments: {
    question: "Compare React and Vue.js for enterprise applications",
    system_prompt: "You are a senior software architect. Provide detailed technical comparisons with pros and cons.",
    max_tokens: 2048
  }
});

# cURL
curl -X POST https://api.joinreeva.com/mcp/server_YOUR_ID \
  -H "Authorization: Bearer mcpk_your_key" \
  -H "Content-Type: application/json" \
  -d '{
    "jsonrpc": "2.0",
    "method": "tools/call",
    "params": {
      "name": "Perplexity_Ask",
      "arguments": {
        "question": "What is MCP (Model Context Protocol)?",
        "max_tokens": 1500
      }
    },
    "id": 1
  }'

Use Cases:

General research questions
Technical explanations
Current events and news
Fact-checking and verification
Summarizing complex topics

Grounded Search

Web search with domain filtering and date constraints.

Tool ID: Perplexity_Grounded_Search
Credit Cost: 3 credits

Parameters:

messages (array, required): Chat messages with role and content
- Include system and user messages as needed
model (string, optional): Model to use
- Default: "sonar-pro"
max_tokens (integer, optional): Maximum tokens
- Default: 1024
temperature (number, optional): Sampling temperature
- Default: 0.2
search_domain_filter (array, optional): Allowlist or denylist of domains
- Maximum: 20 domains
- Allowlist: ["wikipedia.org", "nasa.gov"]
- Denylist: ["-reddit.com", "-twitter.com"] (prefix with -)
search_after_date_filter (string, optional): Only include content after this date
- Format: "1/1/2024" or "2024-01-01"
search_before_date_filter (string, optional): Only include content before this date
search_context_size (integer, optional): Amount of search context to use

Response:

{
  "answer": "According to NASA, the JWST has discovered...",
  "citations": [
    {"url": "https://nasa.gov/jwst", "title": "JWST - NASA"}
  ],
  "model": "sonar-pro"
}

Example Usage:

# Python - Domain-restricted search
response = client.call_tool(
    name="Perplexity_Grounded_Search",
    arguments={
        "messages": [
            {"role": "system", "content": "You are a research assistant."},
            {"role": "user", "content": "What are the latest JWST discoveries?"}
        ],
        "search_domain_filter": ["nasa.gov", "space.com", "esa.int"],
        "search_after_date_filter": "1/1/2024"
    }
)

// TypeScript - Exclude certain domains
const response = await client.callTool({
  name: "Perplexity_Grounded_Search",
  arguments: {
    messages: [
      {role: "user", content: "What is the current state of AI regulation?"}
    ],
    search_domain_filter: ["-reddit.com", "-twitter.com", "-facebook.com"],
    search_after_date_filter: "6/1/2024"
  }
});

Use Cases:

Research from trusted sources only
Exclude social media from results
Time-bound research
Domain-specific investigations

Domain Filter Tips:

Use allowlist for authoritative sources
Use denylist (prefix -) to exclude unreliable sources
Cannot mix allowlist and denylist in same request
Maximum 20 domains per request

arXiv Search

Search arXiv preprints and research papers.

Tool ID: Perplexity_Arxiv_Search
Credit Cost: 3 credits

Parameters:

query (string, required): Research question to search on arXiv
model (string, optional): Model to use
- Default: "sonar-pro"
max_tokens (integer, optional): Maximum tokens
- Default: 1024
temperature (number, optional): Sampling temperature
- Default: 0.2
search_after_date_filter (string, optional): Papers after this date
search_before_date_filter (string, optional): Papers before this date
search_context_size (integer, optional): Amount of context to use

Response:

{
  "answer": "Recent arXiv papers on diffusion models show...",
  "citations": [
    {
      "url": "https://arxiv.org/abs/2301.12345",
      "title": "Advances in Diffusion Models for Image Generation"
    }
  ]
}

Example Usage:

# Python - Search recent ML papers
response = client.call_tool(
    name="Perplexity_Arxiv_Search",
    arguments={
        "query": "State of the art in large language model reasoning",
        "search_after_date_filter": "1/1/2024"
    }
)

print("Research Summary:")
print(response["answer"])

// TypeScript - Physics research
const response = await client.callTool({
  name: "Perplexity_Arxiv_Search",
  arguments: {
    query: "Quantum error correction recent advances",
    max_tokens: 2048
  }
});

Use Cases:

Literature reviews
Finding recent papers
Understanding research trends
Technical deep dives
Pre-print discovery

Search Academic Papers

Search scholarly and peer-reviewed publications.

Tool ID: perplexity_Search_Academic_Papers
Credit Cost: 3 credits

Parameters:

query (string, required): Academic or scholarly research question
model (string, optional): Model to use
- Default: "sonar-pro"
max_tokens (integer, optional): Maximum tokens
- Default: 1024
temperature (number, optional): Sampling temperature
- Default: 0.2
search_after_date_filter (string, optional): Papers after this date
search_before_date_filter (string, optional): Papers before this date

Response:

{
  "answer": "Peer-reviewed research indicates that...",
  "citations": [
    {
      "url": "https://doi.org/10.1234/example",
      "title": "Journal Article Title"
    }
  ]
}

Example Usage:

# Python - Academic research
response = client.call_tool(
    name="perplexity_Search_Academic_Papers",
    arguments={
        "query": "What is the effectiveness of cognitive behavioral therapy for anxiety?",
        "search_after_date_filter": "1/1/2020"
    }
)

// TypeScript - Medical research
const response = await client.callTool({
  name: "perplexity_Search_Academic_Papers",
  arguments: {
    query: "Latest clinical trials on mRNA vaccines",
    max_tokens: 2048
  }
});

Use Cases:

Medical and clinical research
Social science studies
Peer-reviewed evidence
Systematic reviews
Academic citations

Difference from arXiv Search:

Academic Papers: Broader scholarly sources including journals
arXiv Search: Specifically arXiv preprints (CS, physics, math, etc.)

Search Financial Data

Search SEC filings and official financial documents.

Tool ID: perplexity_Search_Financial_Data
Credit Cost: 3 credits

Parameters:

query (string, required): Financial question or SEC filings request
model (string, optional): Model to use
- Default: "sonar-pro"
max_tokens (integer, optional): Maximum tokens
- Default: 1024
temperature (number, optional): Sampling temperature
- Default: 0.2
search_after_date_filter (string, optional): Filings after this date
search_before_date_filter (string, optional): Filings before this date

Response:

{
  "answer": "According to Apple Inc.'s latest 10-K filing...",
  "citations": [
    {
      "url": "https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0000320193",
      "title": "Apple Inc. SEC Filings"
    }
  ]
}

Example Usage:

# Python - Analyze company filings
response = client.call_tool(
    name="perplexity_Search_Financial_Data",
    arguments={
        "query": "What are the key risk factors in Tesla's latest 10-K?",
        "search_after_date_filter": "1/1/2024"
    }
)

// TypeScript - Financial research
const response = await client.callTool({
  name: "perplexity_Search_Financial_Data",
  arguments: {
    query: "Compare revenue growth of FAANG companies in latest quarterly reports"
  }
});

SEC Filing Types:

Filing	Description
10-K	Annual report
10-Q	Quarterly report
8-K	Current report (material events)
S-1	IPO registration
DEF 14A	Proxy statement

Use Cases:

Company financial analysis
Risk factor assessment
Earnings and revenue research
Regulatory compliance tracking
Investment research

Wikipedia Search

Search and summarize Wikipedia articles.

Tool ID: Perplexity_Wikipedia_Search
Credit Cost: 3 credits

Parameters:

query (string, required): Question to answer using Wikipedia sources
model (string, optional): Model to use
- Default: "sonar-pro"
max_tokens (integer, optional): Maximum tokens
- Default: 1024
temperature (number, optional): Sampling temperature
- Default: 0.2
search_after_date_filter (string, optional): Content after this date
search_before_date_filter (string, optional): Content before this date

Response:

{
  "answer": "The James Webb Space Telescope, according to Wikipedia...",
  "citations": [
    {
      "url": "https://en.wikipedia.org/wiki/James_Webb_Space_Telescope",
      "title": "James Webb Space Telescope - Wikipedia"
    }
  ]
}

Example Usage:

# Python - Encyclopedia lookup
response = client.call_tool(
    name="Perplexity_Wikipedia_Search",
    arguments={
        "query": "What is the history of the Turing Award?"
    }
)

// TypeScript - Historical research
const response = await client.callTool({
  name: "Perplexity_Wikipedia_Search",
  arguments: {
    query: "Timeline of major events in World War II",
    max_tokens: 2048
  }
});

Use Cases:

Encyclopedia lookups
Historical information
General knowledge queries
Background research
Quick fact-finding

Common Patterns

Research Pipeline

# Multi-source research workflow
def comprehensive_research(topic):
    # Step 1: General overview
    overview = client.call_tool(
        name="Perplexity_Ask",
        arguments={
            "question": f"Give me a comprehensive overview of {topic}",
            "max_tokens": 2000
        }
    )
    
    # Step 2: Academic sources
    academic = client.call_tool(
        name="perplexity_Search_Academic_Papers",
        arguments={
            "query": f"Recent peer-reviewed research on {topic}",
            "search_after_date_filter": "1/1/2023"
        }
    )
    
    # Step 3: arXiv preprints
    arxiv = client.call_tool(
        name="Perplexity_Arxiv_Search",
        arguments={
            "query": f"Latest research papers on {topic}",
            "search_after_date_filter": "1/1/2024"
        }
    )
    
    return {
        "overview": overview["answer"],
        "academic_research": academic["answer"],
        "latest_preprints": arxiv["answer"],
        "all_citations": (
            overview.get("citations", []) + 
            academic.get("citations", []) + 
            arxiv.get("citations", [])
        )
    }

Financial Analysis

# Analyze company financials
def analyze_company(ticker):
    # Get financial data
    financials = client.call_tool(
        name="perplexity_Search_Financial_Data",
        arguments={
            "query": f"Summarize {ticker} latest 10-K filing key metrics and risk factors",
            "max_tokens": 2000
        }
    )
    
    # Get recent news (excluding social media)
    news = client.call_tool(
        name="Perplexity_Grounded_Search",
        arguments={
            "messages": [
                {"role": "user", "content": f"What is the latest news about {ticker}?"}
            ],
            "search_domain_filter": ["-twitter.com", "-reddit.com"],
            "search_after_date_filter": "1/1/2024"
        }
    )
    
    return {
        "financial_summary": financials["answer"],
        "recent_news": news["answer"]
    }

Fact-Checking

# Verify claims with trusted sources
def fact_check(claim):
    # Check Wikipedia
    wiki = client.call_tool(
        name="Perplexity_Wikipedia_Search",
        arguments={"query": claim}
    )
    
    # Check academic sources
    academic = client.call_tool(
        name="perplexity_Search_Academic_Papers",
        arguments={"query": claim}
    )
    
    # Check trusted news sources
    news = client.call_tool(
        name="Perplexity_Grounded_Search",
        arguments={
            "messages": [{"role": "user", "content": claim}],
            "search_domain_filter": [
                "reuters.com", "apnews.com", "bbc.com", "npr.org"
            ]
        }
    )
    
    return {
        "wikipedia": wiki["answer"],
        "academic": academic["answer"],
        "news": news["answer"]
    }

Best Practices

Query Formulation

Be specific and clear in your questions
Include context for better answers
Use appropriate specialized search for domain-specific queries
Frame questions to get actionable answers

Model Settings

Lower temperature (0.1-0.3) for factual queries
Higher temperature (0.5-0.7) for creative tasks
Increase max_tokens for comprehensive answers
Use system_prompt to guide response format

Source Selection

Use domain filters for authoritative sources
Exclude social media for fact-based research
Use date filters for current information
Match tool to research type (academic, financial, etc.)

Cost Management

Each query costs 3 credits
Combine related questions when possible
Use appropriate max_tokens limits
Cache results for repeated queries

Troubleshooting

"No relevant results" Error

Cause: Query too narrow or no matching content

Solutions:

Broaden your search query
Remove date filters temporarily
Try different phrasing
Use a less restrictive domain filter

Incomplete Answers

Cause: max_tokens limit reached

Solutions:

Increase max_tokens parameter
Ask more focused questions
Break complex queries into parts

Irrelevant Citations

Cause: Query matches tangential content

Solutions:

Be more specific in your question
Use domain filters to restrict sources
Add context in system prompt
Use specialized search tools

Rate Limiting

Cause: Too many requests

Solutions:

Add delays between requests
Batch related queries
Check Perplexity API limits
Monitor usage in dashboard

Web Search - Basic web search without AI synthesis
Web Scraper - Extract content from URLs
Firecrawl - Advanced web scraping
Notion - Store research findings

Overview​

Key Features​

Setup​

Get Perplexity API Key​

Add to Reeva​

Available Tools​

Perplexity Ask​

Grounded Search​

arXiv Search​

Search Academic Papers​

Search Financial Data​

Wikipedia Search​

Common Patterns​

Research Pipeline​

Financial Analysis​

Fact-Checking​

Best Practices​

Query Formulation​

Model Settings​

Source Selection​

Cost Management​

Troubleshooting​

"No relevant results" Error​

Incomplete Answers​

Irrelevant Citations​

Rate Limiting​

Related Tools​

See Also​

Overview

Key Features

Setup

Get Perplexity API Key

Add to Reeva

Available Tools

Perplexity Ask

Grounded Search

arXiv Search

Search Academic Papers

Search Financial Data

Wikipedia Search

Common Patterns

Research Pipeline

Financial Analysis

Fact-Checking

Best Practices

Query Formulation

Model Settings

Source Selection

Cost Management

Troubleshooting

"No relevant results" Error

Incomplete Answers

Irrelevant Citations

Rate Limiting

Related Tools

See Also