Perplexity MCP Zerver

https://github.com/wysh3/perplexity-mcp-zerver
Admin
A minimalist research server implementing MCP to deliver AI-powered research capabilities through...
Tokens:4,207
Snippets:17
Trust Score:5.7
License:GPL-3.0
Update:11 months ago
Show doc for...
Tokens:
Raw
### MCP Server Installation and Setup

Source: https://context7.com/wysh3/perplexity-mcp-zerver/llms.txt

Complete installation process for the MCP server including repository cloning, dependency installation, building, and server startup for testing purposes.

```Bash
# Clone and install
git clone https://github.com/wysh3/perplexity-mcp-zerver.git
cd perplexity-mcp-zerver
bun install
bun run build

# Start the server (for testing)
bun run start
```

```JSON
{
  "mcpServers": {
    "perplexity-server": {
      "command": "bun",
      "args": ["/absolute/path/to/build/main.js"],
      "timeout": 300
    }
  }
}
```

--------------------------------

### Install and Build Perplexity MCP Zerver (Bash)

Source: https://github.com/wysh3/perplexity-mcp-zerver/blob/main/README.md

Instructions for cloning the repository, installing dependencies using Bun, and building the project. Assumes Bun runtime and Node.js 18+ are prerequisites.

```bash
git clone https://github.com/wysh3/perplexity-mcp-zerver.git
cd perplexity-mcp-zerver
bun install
bun run build
```

--------------------------------

### Retrieve Technical Documentation in TypeScript

Source: https://context7.com/wysh3/perplexity-mcp-zerver/llms.txt

Examples of fetching documentation with contextual requirements. Returns formatted documentation with examples and official source URLs.

```typescript
const docRequest = {
  method: "tools/call",
  params: {
    name: "get_documentation",
    arguments: {
      query: "React useEffect hook",
      context: "include cleanup and dependency array examples"
    }
  }
};
```

```typescript
const versionSpecificDoc = {
  method: "tools/call",
  params: {
    name: "get_documentation",
    arguments: {
      query: "TypeScript 5.0 decorators",
      context: "focus on stage 3 decorator syntax changes"
    }
  }
};
```

--------------------------------

### Initialize Perplexity MCP Zerver in TypeScript

Source: https://context7.com/wysh3/perplexity-mcp-zerver/llms.txt

Sets up the PerplexityServer instance and runs it. This is the entry point for using all MCP tools provided by the server.

```typescript
import { PerplexityServer } from "./server/PerplexityServer.js";

// Initialize the server
const server = new PerplexityServer();
await server.run();
```

--------------------------------

### Initialize Perplexity Server with TypeScript

Source: https://context7.com/wysh3/perplexity-mcp-zerver/llms.txt

Demonstrates standard and custom initialization of the PerplexityServer in TypeScript. Standard initialization uses default modules, while custom initialization allows injecting specific instances of DatabaseManager, BrowserManager, and SearchEngine. It also shows how to access these components after the server is initialized.

```typescript
import { PerplexityServer } from "./server/PerplexityServer.js";
import { BrowserManager } from "./server/modules/BrowserManager.js";
import { SearchEngine } from "./server/modules/SearchEngine.js";
import { DatabaseManager } from "./server/modules/DatabaseManager.js";

// Standard initialization
const server = new PerplexityServer();
await server.run();

// Custom initialization with dependency injection
const customDb = new DatabaseManager("./custom-path/chat.db");
const customBrowser = new BrowserManager();
const customSearch = new SearchEngine(customBrowser);

const customServer = new PerplexityServer({
  databaseManager: customDb,
  browserManager: customBrowser,
  searchEngine: customSearch
});

await customServer.run();

// Access components for testing or advanced usage
const browserManager = customServer.getBrowserManager();
const searchEngine = customServer.getSearchEngine();
const databaseManager = customServer.getDatabaseManager();

```

--------------------------------

### Configure Perplexity MCP Zerver in MCP (JSON)

Source: https://github.com/wysh3/perplexity-mcp-zerver/blob/main/README.md

Configuration snippet for integrating the Perplexity MCP Zerver into an MCP client. Specifies the command to run the server and its arguments, along with a timeout setting. Requires the absolute path to the built JavaScript file.

```json
{
  "mcpServers": {
    "perplexity-server": {
      "command": "bun",
      "args": ["/absolute/path/to/build/main.js"],
      "timeout": 300
    }
  }
}
```

--------------------------------

### Find APIs - API Discovery and Comparison

Source: https://context7.com/wysh3/perplexity-mcp-zerver/llms.txt

This endpoint allows you to discover and compare APIs based on specific requirements. It provides information about features, pricing, SDKs, and integration complexity.

```APIDOC
## POST /tools/call (find_apis)

### Description
Discovers and compares APIs based on specific requirements, providing information about features, pricing, SDKs, and integration complexity.

### Method
POST

### Endpoint
/tools/call

### Parameters
#### Request Body
- **name** (string) - Required - The name of the tool to call, which is `find_apis` for this function.
- **arguments** (object) - Required - The arguments for the `find_apis` tool.
  - **requirement** (string) - Required - The primary requirement for the API search (e.g., "payment processing").
  - **context** (string) - Optional - Additional context or constraints for the search (e.g., "needs Stripe alternative with crypto support").

### Request Example
```json
{
  "method": "tools/call",
  "params": {
    "name": "find_apis",
    "arguments": {
      "requirement": "payment processing",
      "context": "needs Stripe alternative with crypto support"
    }
  }
}
```

### Response
#### Success Response (200)
- **response** (string) - A detailed response containing API recommendations and comparisons, including fees, crypto support, SDKs, and official URLs.

#### Response Example
```json
{
  "response": "Here are payment processing alternatives to Stripe:\n\n1. PayPal Commerce Platform\n   - Fees: 2.9% + $0.30 per transaction\n   - Crypto: Yes (via PayPal Crypto)\n   - SDK: Node.js, Python, PHP\n\n2. Square Payments API\n   - Fees: 2.6% + $0.10\n   - Crypto: Limited support\n   - SDK: Multiple languages\n\nOfficial URL(s):\n- https://developer.paypal.com\n- https://developer.squareup.com"
}
```
```

--------------------------------

### API Discovery and Comparison Tool

Source: https://context7.com/wysh3/perplexity-mcp-zerver/llms.txt

Discovers and compares APIs based on specific requirements, providing information about features, pricing, SDKs, and integration complexity. Supports searching for payment processing APIs, image recognition services, and other API categories with detailed comparisons.

```TypeScript
// Find payment processing APIs
const apiRequest = {
  method: "tools/call",
  params: {
    name: "find_apis",
    arguments: {
      requirement: "payment processing",
      context: "needs Stripe alternative with crypto support"
    }
  }
};

// Response with API recommendations and comparisons
{
  "response": "Here are payment processing alternatives to Stripe:\n\n1. PayPal Commerce Platform\n   - Fees: 2.9% + $0.30 per transaction\n   - Crypto: Yes (via PayPal Crypto)\n   - SDK: Node.js, Python, PHP\n\n2. Square Payments API\n   - Fees: 2.6% + $0.10\n   - Crypto: Limited support\n   - SDK: Multiple languages\n\nOfficial URL(s):\n- https://developer.paypal.com\n- https://developer.squareup.com"
}

// Find free-tier APIs
const freeTierSearch = {
  method: "tools/call",
  params: {
    name: "find_apis",
    arguments: {
      requirement: "image recognition and classification",
      context: "prefer free tier options with at least 1000 requests/month"
    }
  }
};
```

--------------------------------

### Perform Web Search with Perplexity AI in TypeScript

Source: https://context7.com/wysh3/perplexity-mcp-zerver/llms.txt

Demonstrates how to make search requests with configurable detail levels (brief, normal, detailed) and streaming options. Returns search results with source URLs.

```typescript
const searchRequest = {
  method: "tools/call",
  params: {
    name: "search",
    arguments: {
      query: "What are the latest advancements in quantum computing?",
      detail_level: "detailed",
      stream: false
    }
  }
};
```

```typescript
const briefSearch = {
  method: "tools/call",
  params: {
    name: "search",
    arguments: {
      query: "Python async/await syntax",
      detail_level: "brief"
    }
  }
};
```

--------------------------------

### Legacy Code Analysis and Migration Tool

Source: https://context7.com/wysh3/perplexity-mcp-zerver/llms.txt

Analyzes code snippets to identify deprecated patterns, suggests modern alternatives, and provides migration guidance for technology upgrades. Supports multiple technologies including React and Node.js with specific deprecation warnings and recommended replacements.

```TypeScript
// Check React lifecycle method
const deprecationCheck = {
  method: "tools/call",
  params: {
    name: "check_deprecated_code",
    arguments: {
      code: "componentWillMount() {\n  this.setState({ loading: true });\n}",
      technology: "React 18"
    }
  }
};

// Response with deprecation analysis and alternatives
{
  "response": "⚠️ DEPRECATED: componentWillMount() was deprecated in React 16.3 and removed in React 17.\n\nRecommended alternatives:\n1. Use constructor() for initialization\n2. Use componentDidMount() for side effects\n3. Use React Hooks (useEffect) in functional components\n\nMigration example:\nuseEffect(() => {\n  setLoading(true);\n}, []);\n\nOfficial URL(s):\n- https://react.dev/reference/react/Component#unsafe_componentwillmount"
}

// Check Node.js APIs
const nodeDeprecation = {
  method: "tools/call",
  params: {
    name: "check_deprecated_code",
    arguments: {
      code: "const crypto = require('crypto');\ncrypto.createCipher('aes192', password);",
      technology: "Node.js 20"
    }
  }
};
```

--------------------------------

### Check Deprecated Code - Legacy Code Analysis

Source: https://context7.com/wysh3/perplexity-mcp-zerver/llms.txt

Analyzes code snippets to identify deprecated patterns, suggests modern alternatives, and provides migration guidance for technology upgrades.

```APIDOC
## POST /tools/call (check_deprecated_code)

### Description
Analyzes code snippets to identify deprecated patterns, suggests modern alternatives, and provides migration guidance for technology upgrades.

### Method
POST

### Endpoint
/tools/call

### Parameters
#### Request Body
- **name** (string) - Required - The name of the tool to call, which is `check_deprecated_code` for this function.
- **arguments** (object) - Required - The arguments for the `check_deprecated_code` tool.
  - **code** (string) - Required - The code snippet to analyze for deprecation.
  - **technology** (string) - Required - The technology and version to check against (e.g., "React 18", "Node.js 20").

### Request Example
```json
{
  "method": "tools/call",
  "params": {
    "name": "check_deprecated_code",
    "arguments": {
      "code": "componentWillMount() {\n  this.setState({ loading: true });\n}",
      "technology": "React 18"
    }
  }
}
```

### Response
#### Success Response (200)
- **response** (string) - A detailed response including deprecation analysis, recommended alternatives, migration examples, and official URLs.

#### Response Example
```json
{
  "response": "⚠️ DEPRECATED: componentWillMount() was deprecated in React 16.3 and removed in React 17.\n\nRecommended alternatives:\n1. Use constructor() for initialization\n2. Use componentDidMount() for side effects\n3. Use React Hooks (useEffect) in functional components\n\nMigration example:\nuseEffect(() => {\n  setLoading(true);\n}, []);\n\nOfficial URL(s):\n- https://react.dev/reference/react/Component#unsafe_componentwillmount"
}
```
```

--------------------------------

### Chat with Context in Perplexity MCP Zerver (TypeScript)

Source: https://context7.com/wysh3/perplexity-mcp-zerver/llms.txt

Shows how to initiate and continue conversations with context history. Responses include chat_id for maintaining session state across multiple requests.

```typescript
const chatRequest = {
  method: "tools/call",
  params: {
    name: "chat_perplexity",
    arguments: {
      message: "Explain how neural networks learn from data"
    }
  }
};
```

```typescript
const followUpRequest = {
  method: "tools/call",
  params: {
    name: "chat_perplexity",
    arguments: {
      message: "How does this differ from traditional machine learning?",
      chat_id: "123e4567-e89b-12d3-a456-426614174000"
    }
  }
};
```

--------------------------------

### Web Content Extraction and Mining Tool

Source: https://context7.com/wysh3/perplexity-mcp-zerver/llms.txt

Extracts clean article text from web pages using Mozilla's Readability library, with special handling for GitHub repositories via gitingest.com integration. Supports recursive exploration with configurable depth and handles partial success scenarios.

```TypeScript
// Extract article content
const extractRequest = {
  method: "tools/call",
  params: {
    name: "extract_url_content",
    arguments: {
      url: "https://example.com/blog/machine-learning-guide",
      depth: 1
    }
  }
};

// Response with structured content
{
  "status": "Success",
  "rootUrl": "https://example.com/blog/machine-learning-guide",
  "explorationDepth": 1,
  "pagesExplored": 1,
  "content": [
    {
      "url": "https://example.com/blog/machine-learning-guide",
      "title": "A Comprehensive Guide to Machine Learning",
      "textContent": "Machine learning is a subset of artificial intelligence that enables systems to learn and improve..."
    }
  ]
}

// Recursive exploration with depth
const recursiveExtract = {
  method: "tools/call",
  params: {
    name: "extract_url_content",
    arguments: {
      url: "https://docs.example.com/getting-started",
      depth: 3
    }
  }
};

// GitHub repository extraction
const githubExtract = {
  method: "tools/call",
  params: {
    name: "extract_url_content",
    arguments: {
      url: "https://github.com/facebook/react",
      depth: 1
    }
  }
};

// Handles partial success scenarios
{
  "status": "SuccessWithPartial",
  "message": "Fetched 8/10 pages successfully. Some pages failed or timed out.",
  "rootUrl": "https://docs.example.com/getting-started",
  "explorationDepth": 3,
  "pagesExplored": 8,
  "content": [
    // ... successfully extracted pages
  ]
}
```

--------------------------------

### Extract URL Content - Web Content Extraction

Source: https://context7.com/wysh3/perplexity-mcp-zerver/llms.txt

Extracts clean article text from web pages using Mozilla's Readability library, with special handling for GitHub repositories via gitingest.com integration.

```APIDOC
## POST /tools/call (extract_url_content)

### Description
Extracts clean article text from web pages using Mozilla's Readability library, with special handling for GitHub repositories via gitingest.com integration.

### Method
POST

### Endpoint
/tools/call

### Parameters
#### Request Body
- **name** (string) - Required - The name of the tool to call, which is `extract_url_content` for this function.
- **arguments** (object) - Required - The arguments for the `extract_url_content` tool.
  - **url** (string) - Required - The URL of the web page to extract content from.
  - **depth** (integer) - Optional - The depth for recursive exploration of links. Defaults to 1.

### Request Example
```json
{
  "method": "tools/call",
  "params": {
    "name": "extract_url_content",
    "arguments": {
      "url": "https://example.com/blog/machine-learning-guide",
      "depth": 1
    }
  }
}
```

### Response
#### Success Response (200)
- **status** (string) - Indicates the success status of the operation (e.g., "Success", "SuccessWithPartial").
- **rootUrl** (string) - The original URL that was requested.
- **explorationDepth** (integer) - The depth that was explored.
- **pagesExplored** (integer) - The number of pages successfully explored.
- **content** (array) - An array of objects, where each object represents an extracted page.
  - **url** (string) - The URL of the extracted page.
  - **title** (string) - The title of the extracted page.
  - **textContent** (string) - The cleaned text content of the page.

#### Response Example (Success)
```json
{
  "status": "Success",
  "rootUrl": "https://example.com/blog/machine-learning-guide",
  "explorationDepth": 1,
  "pagesExplored": 1,
  "content": [
    {
      "url": "https://example.com/blog/machine-learning-guide",
      "title": "A Comprehensive Guide to Machine Learning",
      "textContent": "Machine learning is a subset of artificial intelligence that enables systems to learn and improve..."
    }
  ]
}
```

#### Response Example (Partial Success)
```json
{
  "status": "SuccessWithPartial",
  "message": "Fetched 8/10 pages successfully. Some pages failed or timed out.",
  "rootUrl": "https://docs.example.com/getting-started",
  "explorationDepth": 3,
  "pagesExplored": 8,
  "content": [
    // ... successfully extracted pages
  ]
}
```
```
Tokens:
Raw