Exa MCP Server vs Perplexity MCP Server: both are research tools on Scriptiii AI Tools. Exa MCP Server pricing: Free. Free to use. (reported by the maker) Perplexity MCP Server pricing: Free. Free to use. (reported by the maker) Both have 0 upvotes.
| Field | Exa MCP Server | Perplexity MCP Server |
|---|---|---|
| Page | Exa MCP Server | Perplexity MCP Server |
| Pricing (maker-reported) | Free. Free to use. | Free. Free to use. |
| Upvotes | 0 | 0 |
| Best for | ai-agents, mcp-server, search | ai-agents, mcp-server, search |
| Tagline | MCP server for Exa web search and crawling, available as a hosted endpoint or local package. | Official MCP server for the Perplexity API, offering web-grounded search and answers. |
| Website | Visit Exa MCP Server | Visit Perplexity MCP Server |
Pricing and pricing details come from the maker's submission. They are not independently verified.
Connect AI agents to Exa for web search, content fetching, and multi-step research.
Documentation · NPM Package · Get your API key
Connect to Exa's hosted MCP server:
https://mcp.exa.ai/mcp
Or use a plugin when your client supports one.
This repository is an Agent Plugin. Install it with any compatible client.
Install from the Claude Plugin Marketplace, or run:
claude plugin install exa@claude-plugins-official
Install via Plugins in ChatGPT, or run:
codex mcp add exa --url https://mcp.exa.ai/mcp
Most clients can be configured manually with the standard mcpServers shape:
{
"mcpServers": {
"exa": {
"type": "streamable-http",
"url": "https://mcp.exa.ai/mcp",
}
}
}
Exa MCP works with most other clients, point them at https://mcp.exa.ai/mcp.
| Client | Where to add it |
|---|---|
| Kiro | Use the Kiro power, or add manually to ~/.kiro/settings/mcp.json |
| LM Studio | Add to LM Studio, or add manually to mcp.json |
| Replit | Add to Replit |
| Grok Build | /marketplace → install Exa, then /mcp to sign in |
| Gemini CLI | Add manually to ~/.gemini/settings.json |
| OpenCode | Add manually to opencode.json |
| Windsurf | Add manually to ~/.codeium/windsurf/mcp_config.json |
| Google Antigravity | Add manually to mcp_config.json |
| Zed | Add manually to settings.json under context_servers |
| Warp | Settings → Agents → MCP servers |
| v0 by Vercel | Settings → MCP connections |
| Tool | Description |
|---|---|
web_search_exa | Search the web for any topic and get clean, ready-to-use content |
web_fetch_exa | Read a webpage's full content as clean markdown from one or more URLs |
tools parameter)| Tool | Description |
|---|---|
agent_run | Run an Exa Agent for multi-step research, list-building, enrichment, and structured output |
web_search_advanced_exa | Advanced search with filters, domains, dates, highlights, summaries, and subpage crawling |
Enable tools by appending them to the MCP URL (this will replace the defaults, so include all you want):
https://mcp.exa.ai/mcp?tools=web_search_advanced_exa
https://mcp.exa.ai/mcp?tools=web_search_exa,web_fetch_exa,agent_run
Exa Agent requires authentication (OAuth or an API key).
Add dynamicHighlights=true to the MCP URL to make web_search_exa and web_search_advanced_exa return Dynamic Highlights. These are highlights chosen across all results under one shared budget, not a fixed budget per URL, so less text reaches your agent's context. When the server runs locally, set DYNAMIC_HIGHLIGHTS=true instead.
https://mcp.exa.ai/mcp?dynamicHighlights=true
Skills live in skills/ and load with Agent Plugin installs. The Claude plugin in plugins/exa/ ships a copy of them; npm test fails when the copy drifts. Each package version bump on main publishes plugins/exa/ to the root of the claude-plugin branch, which Claude's plugin directory reads. To publish a plugin-only change between releases, run the Publish Claude plugin workflow manually.
| Skill | Path | Use when |
|---|---|---|
search | skills/search/ | Deep research, lead gen, competitive analysis, multi-step web investigation |
exa-agent | skills/exa-agent/ | Exa Agent runs, enrichment, structured output, Connect providers |
Invoke from your client's skill UI (or /skill-name where supported). MCP-only setups still get the tools; skills add orchestration on top.
The hosted MCP server works anonymously with rate limits. For higher limits and access to Exa Agent, use either OAuth or an API key.
OAuth is preferred: most clients prompt you to sign in to Exa. To force the login flow (useful for shared connectors and plugins), use https://mcp.exa.ai/mcp?login or https://mcp.exa.ai/mcp/oauth.
If you prefer, you can get an API key from the dashboard and pass it on the URL as ?exaApiKey=…. You can also send it as a Authorization: Bearer … header or an x-api-key header.
Built with ❤️ by Exa
The official MCP server implementation for the Perplexity API Platform, providing AI assistants with real-time web search, reasoning, and research capabilities through the Agent API and the Search API.
The remote MCP server is hosted by Perplexity and is the easiest way to get started: same tools, nothing to install or update. The Cursor and VS Code buttons at the top of this page connect to it with one click. If your MCP client does not support remote servers yet, skip to the local server setup below. Connect over Streamable HTTP with your :
https://api.perplexity.ai/mcp
For Claude Code:
claude mcp add --transport http perplexity https://api.perplexity.ai/mcp --header "Authorization: Bearer YOUR_API_KEY"
See the MCP integration docs for manual Cursor/VS Code configuration, usage from the Anthropic API, and setup for other clients.
your_key_here in the configurations below with your API keyPERPLEXITY_TIMEOUT_MS=600000 (default: 5 minutes)PERPLEXITY_BASE_URL=https://your-custom-url.com (default: https://api.perplexity.ai)PERPLEXITY_LOG_LEVEL=DEBUG|INFO|WARN|ERROR (default: ERROR)claude mcp add perplexity --env PERPLEXITY_API_KEY="your_key_here" -- npx -y @perplexity-ai/mcp-server
Or install via plugin:
export PERPLEXITY_API_KEY="your_key_here"
claude
## Then run: /plugin marketplace add perplexityai/modelcontextprotocol
## Then run: /plugin install perplexity
codex mcp add perplexity --env PERPLEXITY_API_KEY="your_key_here" -- npx -y @perplexity-ai/mcp-server
This repository is packaged as an Agent Plugin, so clients that support the standard can install it directly from this repository. The Agent Plugins format does not carry secrets, so set the PERPLEXITY_API_KEY environment variable through your client's plugin or MCP settings.
Most clients can be configured manually using the same mcpServers wrapper in their client config (as shown for Cursor). If a client has a different schema, check its docs for the exact wrapper format.
For manual setup, these clients all use the same mcpServers structure:
| Client | Config File |
|---|---|
| Cursor | ~/.cursor/mcp.json |
| Claude Desktop | claude_desktop_config.json |
| Kiro | .kiro/settings/mcp.json |
| Windsurf | ~/.codeium/windsurf/mcp_config.json |
| VS Code | .vscode/mcp.json |
{
"mcpServers": {
"perplexity": {
"command": "npx",
"args": ["-y", "@perplexity-ai/mcp-server"],
"env": {
"PERPLEXITY_API_KEY": "your_key_here"
}
}
}
}
If you are running this server at work—especially behind a company firewall or proxy—you may need to tell the program how to send its internet traffic through your network's proxy. Follow these steps:
1. Get your proxy details
2. Set the proxy environment variable
The easiest and most reliable way for Perplexity MCP is to use PERPLEXITY_PROXY. For example:
export PERPLEXITY_PROXY=https://your-proxy-host:8080
If your proxy needs a username and password, use:
export PERPLEXITY_PROXY=https://username:password@your-proxy-host:8080
3. Alternate: Standard environment variables
If you'd rather use the standard variables, we support HTTPS_PROXY and HTTP_PROXY.
[!NOTE]
The server checks proxy settings in this order:PERPLEXITY_PROXY→HTTPS_PROXY→HTTP_PROXY. If none are set, it connects directly to the internet.
URLs must includehttps://. Typical ports are8080,3128, and80.
For cloud or shared deployments, run the server in HTTP mode.
| Variable | Description | Default |
|---|---|---|
PERPLEXITY_API_KEY | Your Perplexity API key | Required |
PERPLEXITY_BASE_URL | Custom base URL for API requests | https://api.perplexity.ai |
PORT | HTTP server port | 8080 |
BIND_ADDRESS | Network interface to bind to. Defaults to loopback. Set to 0.0.0.0 to expose on all interfaces. | 127.0.0.1 |
ALLOWED_ORIGINS | CORS origins (comma-separated). Defaults to empty (no cross-origin browser requests). Set to an explicit allowlist (e.g. https://app.example.com) or to * to allow any origin. | (empty) |
ALLOWED_HOSTS | Additional Host header values to accept (comma-separated). Loopback hosts on PORT are always allowed. Add the public hostname when binding to 0.0.0.0. | (loopback only) |
docker build -t perplexity-mcp-server .
docker run -p 8080:8080 -e PERPLEXITY_API_KEY=your_key_here perplexity-mcp-server
export PERPLEXITY_API_KEY=your_key_here
npm install && npm run build && npm run start:http
The server will be accessible at http://localhost:8080/mcp
Search results include publication and last-updated dates when available. Missing or null dates are omitted from the formatted output.
Direct web search using the Perplexity Search API. Returns ranked search results with metadata, perfect for finding current information. Supports recency filters (search_recency_filter), domain restrictions (search_domain_filter), and Fast Search (search_type: "fast") for lower latency and cost.
General-purpose conversational AI with real-time web search, backed by the Agent API fast preset. Great for quick questions and everyday searches.
Deep, comprehensive research backed by the Agent API high preset. Ideal for thorough analysis and detailed reports. Runs can take minutes; the server streams the run and reports progress to clients that request it.
Advanced reasoning and problem-solving backed by the Agent API medium preset. Perfect for complex analytical tasks.
perplexity_ask and perplexity_reason accept optional positive-integer max_tokens (total search context, up to 120,000) and max_tokens_per_page (context per result page, up to 20,000). These Agent API budgets control retrieved search content, not generated answer length. When omitted, the preset's search configuration remains in effect.
The legacy search_context_size argument is still accepted, but is translated locally into explicit budgets rather than sent to the Agent API:
| Legacy size | Total tokens | Tokens per page |
|---|---|---|
low | 2000 | 512 |
medium | 4000 | 1024 |
high | 8000 | 2048 |
This compatibility policy is not an exact equivalent of Sonar's named sizes; the Agent API migration guide has no named-size equivalent. Each explicit budget overrides its corresponding legacy default.
[!NOTE]
Presets are managed configurations (model, search setup, step budget) that Perplexity keeps tuned over time; see the presets guide. Earlier versions of this server called the legacysonar-pro,sonar-reasoning-pro, andsonar-deep-researchmodels and acceptedstrip_thinking/reasoning_effortparameters. Those parameters are no longer part of the tool schemas and are ignored if sent; the Agent API produces no<think>tags.
The package also exports the server factory for embedding in your own Node process:
import { createPerplexityServer } from "@perplexity-ai/mcp-server";
// Single-tenant: reads PERPLEXITY_API_KEY from the environment.
const server = createPerplexityServer("my-service");
// Multi-tenant hosts resolve the key per call instead. When a provider is
// set, the environment variable is never consulted, and a provider that
// returns no key fails the call rather than falling back.
const tenantServer = createPerplexityServer("my-service", {
apiKey: () => currentRequestApiKey,
});
Mount the returned server on any MCP transport (stdio, streamable HTTP, in-memory).
PERPLEXITY_API_KEY is set correctlyPERPLEXITY_TIMEOUT_MS to a higher valuePERPLEXITY_PROXY or HTTPS_PROXY setup and ensure api.perplexity.ai isn't blocked by your firewall.npx writes installation messages to stdout. Use npx -yq instead of npx -y to suppress this output.For support, visit community.perplexity.ai or file an issue.