---
name: bu878zo7.nx.link
description: bu878zo7.nx.link exposes a single skill that performs neural semantic search over the web using Exa.ai, returning ranked results with titles, URLs, images, and favicons. It accepts natural-language queries and returns metadata-level results — not full page content.
host: bu878zo7.nx.link
---

# bu878zo7.nx.link

This host wraps Exa.ai's semantic search API, serving agents that need to discover relevant web documents or reference URLs from natural-language queries. It is distinct from keyword-based search in that it uses neural embeddings to match intent, not just terms. It is a single-purpose host with no additional data enrichment or content extraction capabilities.

## When to use this host

Use this host when an agent needs to find relevant web documents, citations, or reference URLs for a natural-language query and keyword search would be insufficient. Do not use it when full page text or document content is required — it returns metadata and URLs only. Do not use it for real-time data feeds, live APIs, or structured database queries. If full content extraction is needed after retrieving URLs, a separate scraping or reader skill from a different host would be required. If keyword-precision search is needed rather than semantic intent matching, a traditional search API host would be more appropriate.

## Capabilities

### Semantic Web Search

Accepts a natural-language query and returns ranked web results including titles, URLs, images, and favicons, using neural semantic matching rather than keyword lookup.

- **`search-exa-semantic`** — Performs a neural semantic search over the web and returns ranked results with titles, URLs, images, and favicons for a natural-language query.

## Skill reference

### `search-exa-semantic`

**Exa.ai Semantic Search** — Performs a neural semantic search over the web and returns ranked results with titles, URLs, images, and favicons for a natural-language query.

*Use when:* Use when an agent needs to retrieve relevant web documents, citations, or reference URLs for a natural-language query, especially when keyword search would be insufficient.

*Not for:* Do not use for real-time data feeds, structured database lookups, or queries requiring full page content extraction — this returns metadata and URLs only, not full document text.

**Inputs:**

- `query` (string, required) — Natural-language search query to run against the semantic index.
- `limit` (integer) — Maximum number of results to return.

**Returns:** Returns a requestId, an array of up to N results each with id/title/url and optional image/favicon, plus searchTime in ms and a costDollars breakdown.

**Example:** `{"query": "solana", "limit": 10}`

---