Bring Neural Search
to Vercel AI SDK
Learn how to connect Exa to Vercel AI SDK and start using 10 AI agent tools in minutes. Fully managed, enterprise secure, and ready to use without writing a single line of code.
What is the Exa MCP Server?
Connect Exa (formerly Metaphor) to any AI agent and simplify how you discover relevant web data, research topics using neural embeddings, and extract cleaned content through natural conversation.
What you can do
- Neural Search — Perform AI-powered searches that understand the meaning and intent of your query, not just keywords.
- Semantic Similarity — Find web pages and articles that are semantically similar to a specific seed URL.
- Content Extraction — Fetch cleaned, readable text and metadata from multiple URLs or page IDs in a single request.
- Research Automation — Discover high-quality sources and analyze web content programmatically via AI.
- Data Insights — retrieve comprehensive metadata for search results to identify top-performing resources.
- Intelligent Discovery — Use 'Autoprompt' to automatically optimize your queries for neural search accuracy.
How it works
1. Subscribe to this server
2. Enter your Exa API Key (found in your account dashboard)
3. Start searching the web for your AI projects from Claude, Cursor, or any MCP client
Who is this for?
- AI Developers & Researchers — quickly find semantically relevant data and clean web content via simple AI commands.
- Content Strategists — discover similar articles and research topics across the web directly from the workspace.
- Knowledge Workers — automate information discovery and extract key insights via the AI assistant.
Built-in capabilities (10)
Get an AI answer
Great for competitive analysis and research. Find similar pages
Find similar pages with content
Extract page contents
Returns titles, URLs, and relevance scores for matching pages. Search the web with AI
Useful for site-specific research or documentation lookups. Search within a domain
Traditional keyword search
Ideal for conceptual queries and research topics. Neural semantic search
Ideal for news and trending topics. Search recent content
Search and extract page content
Why Vercel AI SDK?
The Vercel AI SDK gives every Exa tool full TypeScript type inference, IDE autocomplete, and compile-time error checking. Connect 10 tools through Vinkius and stream results progressively to React, Svelte, or Vue components. works on Edge Functions, Cloudflare Workers, and any Node.js runtime.
- —
TypeScript-first: every MCP tool gets full type inference, IDE autocomplete, and compile-time error checking out of the box
- —
Framework-agnostic core works with Next.js, Nuxt, SvelteKit, or any Node.js runtime. same Exa integration everywhere
- —
Built-in streaming UI primitives let you display Exa tool results progressively in React, Svelte, or Vue components
- —
Edge-compatible: the AI SDK runs on Vercel Edge Functions, Cloudflare Workers, and other edge runtimes for minimal latency
Exa in Vercel AI SDK
Exa and 3,400+ other MCP servers. One platform. One governance layer.
Teams that connect Exa to Vercel AI SDK through Vinkius don't need to source, host, or maintain individual MCP servers. Every tool call runs inside a hardened runtime with credential isolation, DLP, and a signed audit chain.
Raw MCP | Vinkius | |
|---|---|---|
| Server catalog | Find and host yourself | 3,400+ managed |
| Infrastructure | Self-hosted | Sandboxed V8 isolates |
| Credential handling | Plaintext in config | Vault + runtime injection |
| Data loss prevention | None | Configurable DLP policies |
| Kill switch | None | Global instant shutdown |
| Financial circuit breakers | None | Per-server limits + alerts |
| Audit trail | None | Ed25519 signed logs |
| SIEM log streaming | None | Splunk, Datadog, Webhook |
| Honeytokens | None | Canary alerts on leak |
| Custom domains | Not applicable | DNS challenge verified |
| GDPR compliance | Manual effort | Automated purge + export |
Why teams choose Vinkius for Exa in Vercel AI SDK
The Exa MCP Server runs on Vinkius-managed infrastructure inside AWS — a purpose-built runtime with per-request V8 isolates, Ed25519 signed audit chains, and sub-40ms cold starts. All 10 tools execute in hardened sandboxes optimized for native MCP execution.
Your AI agents in Vercel AI SDK only access the data you authorize, with DLP that blocks sensitive information from ever reaching the model, kill switch for instant shutdown, and up to 60% token savings. Enterprise-grade infrastructure, zero maintenance.
* Every MCP server runs on Vinkius-managed infrastructure inside AWS - a purpose-built runtime with per-request V8 isolates, Ed25519 signed audit chains, and sub-40ms cold starts optimized for native MCP execution. See our infrastructure
How Vinkius secures
Exa for Vercel AI SDK
Every tool call from Vercel AI SDK to the Exa MCP Server is protected by DLP redaction, cryptographic audit chains, V8 sandbox isolation, kill switch, and financial circuit breakers.
Frequently asked questions
Can I perform a meaning-based search instead of keywords via AI?
Yes! Use the search tool. Provide your query and enable useAutoprompt. Exa will use its neural embeddings to find links that match the intent of your question.
How do I find articles similar to a specific URL?
Use the find_similar tool. Provide the seed URL, and the agent will retrieve a list of pages that are semantically related to that content.
Is it possible to extract cleaned text from multiple pages via AI?
Absolutely. Use the get_contents query. Provide an array of URLs or Exa IDs, and the agent will return the cleaned text content and metadata for each page.
How does the Vercel AI SDK connect to MCP servers?
Import createMCPClient from @ai-sdk/mcp and pass the server URL. The SDK discovers all tools and provides typed TypeScript interfaces for each one.
Can I use MCP tools in Edge Functions?
Yes. The AI SDK is fully edge-compatible. MCP connections work on Vercel Edge Functions, Cloudflare Workers, and similar runtimes.
Does it support streaming tool results?
Yes. The SDK provides streaming primitives like useChat and streamText that handle tool calls and display results progressively in the UI.
createMCPClient is not a function
Install: npm install @ai-sdk/mcp