superagnt/leverage / skills/core/exa-search
exa-search
VerifiedNeural web search, parsed page contents, find-similar, and agentic Q&A over the live web through the superagnt unified API.
Exa Integration
The Exa integration connects your Exa AI account to superagnt so AI agents and automations can query the live web with neural + keyword search, pull parsed page content (text, highlights, summaries), discover semantically similar URLs, and run agentic Q&A with cited sources. Supports the full set of Exa endpoints — /search, /contents, /findSimilar, and /answer — proxied through your superagnt API key.
Alternative install: the MCP server
If this client speaks MCP, you can connect the workspace server instead of using this skill's curl calls — after the user connects their Exa account in the superagnt app, the same endpoints below are exposed as native MCP tools:
https://mcp.superagnt.com/mcp
The URL publishes full OAuth discovery — an MCP-capable client needs the URL and nothing else (approve once in the browser). Per-client setup lines: https://www.superagnt.com/install
Everything below works on curl-only environments with just the API key.
Capabilities and safety
This skill documents the Exa API surface the user's connected account can reach — which can include destructive operations (updates, deletes) and, where the vendor supports them, actions performed as the user (sending messages, modifying records, changing settings). Two hard rules:
- Confirm before destructive or outbound actions. Never delete, overwrite, or send on the user's behalf without their explicit confirmation in the conversation.
- Stay inside the user's request. Use only the endpoints the task needs; this skill grants no access beyond the Exa connection the user set up themselves.
Prerequisites
- An API key — get one from the superagnt app and export it as
SUPERAGNT_API_KEY. - A connected Exa account — connect it on the Connections page. Calls fail with a clear error until the vendor is connected; that error is the signal to send the user to the connections page, not a bug.
Note: /v1/connections/* calls run on the USER'S Exa credentials —
vendor-side rate limits and billing are theirs, not platform credits.
Authentication
Authorization: Bearer $SUPERAGNT_API_KEY
If the variable is not set, ask the user for their key or point them at https://superagnt.com/r/ch-exa-search-key.
Verify the install (do this first)
curl -s https://api.superagnt.com/v1/credits \
-H 'Authorization: Bearer $SUPERAGNT_API_KEY'
A JSON result proves the key end to end. 401 = bad key; an error naming the connection means the Exa account is not connected yet.
Base URL
https://api.superagnt.com/v1/connections/exa
Available Endpoints
| Method | Path | Summary |
|---|---|---|
POST | /search | Search |
POST | /contents | Get contents |
POST | /findSimilar | Find similar links |
POST | /answer | Generate an answer from search results |
Tool Schemas
The following JSON defines every operation with its parameters. Each tool maps to an API endpoint under the base URL.
[
{
"name": "superagnt_connection_exa_search",
"description": "Search",
"method": "POST",
"path": "/search",
"parameters": {
"type": "object",
"properties": {
"query": {
"type": "string",
"description": "The query string for the search."
},
"additionalQueries": {
"type": "array",
"description": "Additional query variations for deep-search variants. Used alongside the main query for more comprehensive results."
},
"stream": {
"type": "boolean",
"description": "If true, the response is returned as an SSE stream of OpenAI-compatible chat completion chunks."
},
"outputSchema": {
"type": "object",
"description": "JSON schema for synthesized output. When provided, the response includes an output object and output.content matches this schema."
},
"systemPrompt": {
"type": "string",
"description": "Instructions that guide synthesized output and, for deep-search variants, search planning."
},
"type": {
"type": "string",
"description": "Search strategy. `auto` intelligently combines neural and other search methods; `fast` uses streamlined models; `deep-lite` / `deep` / `deep-reasoning` are synthesized-output modes; `instant` is the lowest-latency mode."
},
"category": {
"type": "string",
"description": "A data category to focus on. `people` and `company` support a limited filter set — `startPublishedDate`, `endPublishedDate`, `startCrawlDate`, `endCrawlDate`, and `excludeDomains` are not allowed; `includeDomains` for `people` only accepts LinkedIn domains."
},
"userLocation": {
"type": "string",
"description": "Two-letter ISO country code of the user (e.g. `US`)."
},
"numResults": {
"type": "integer",
"description": "Number of results to return. Max 100 for `neural` and deep-search variants."
},
"includeDomains": {
"type": "array",
"description": "Restrict results to these domains."
},
"excludeDomains": {
"type": "array",
"description": "Exclude these domains from results."
},
"startCrawlDate": {
"type": "string",
"description": "Only include links crawled after this ISO 8601 date."
},
"endCrawlDate": {
"type": "string",
"description": "Only include links crawled before this ISO 8601 date."
},
"startPublishedDate": {
"type": "string",
"description": "Only include links with a published date after this ISO 8601 date."
},
"endPublishedDate": {
"type": "string",
"description": "Only include links with a published date before this ISO 8601 date."
},
"moderation": {
"type": "boolean",
"description": "Enable content moderation to filter unsafe content from results."
},
"contents": {
"type": "object",
"description": "Content enrichment options returned alongside search results. See /contents for the full schema."
}
},
"required": [
"query"
]
}
},
{
"name": "superagnt_connection_exa_getContents",
"description": "Get contents",
"method": "POST",
"path": "/contents",
"parameters": {
"type": "object",
"properties": {
"urls": {
"type": "array",
"description": "URLs to crawl (backwards compatible with `ids`)."
},
"ids": {
"type": "array",
"description": "Document IDs obtained from previous searches."
},
"text": {
"type": "string",
"description": "Boolean or advanced text options object. Object form supports `maxCharacters`, `includeHtmlTags`, `verbosity` (`compact` | `standard` | `full`), `includeSections`, `excludeSections` (header | navigation | banner | body | sidebar | footer | metadata)."
},
"highlights": {
"type": "string",
"description": "Boolean or advanced highlights object. Object form supports `maxCharacters`, `query`."
},
"summary": {
"type": "object",
"description": "LLM-generated summary options."
},
"livecrawl": {
"type": "string",
"description": "Deprecated — use `maxAgeHours`."
},
"livecrawlTimeout": {
"type": "integer",
"description": "livecrawlTimeout"
},
"maxAgeHours": {
"type": "integer",
"description": "Max age of cached content in hours. Positive = livecrawl if older, 0 = always livecrawl, -1 = never livecrawl."
},
"subpages": {
"type": "integer",
"description": "subpages"
},
"subpageTarget": {
"type": "string",
"description": "String or array of strings."
},
"extras": {
"type": "object",
"description": "extras"
}
},
"required": [
"urls"
]
}
},
{
"name": "superagnt_connection_exa_findSimilar",
"description": "Find similar links",
"method": "POST",
"path": "/findSimilar",
"parameters": {
"type": "object",
"properties": {
"url": {
"type": "string",
"description": "The URL to find similar links for."
},
"numResults": {
"type": "integer",
"description": "numResults"
},
"includeDomains": {
"type": "array",
"description": "includeDomains"
},
"excludeDomains": {
"type": "array",
"description": "excludeDomains"
},
"startCrawlDate": {
"type": "string",
"description": "startCrawlDate"
},
"endCrawlDate": {
"type": "string",
"description": "endCrawlDate"
},
"startPublishedDate": {
"type": "string",
"description": "startPublishedDate"
},
"endPublishedDate": {
"type": "string",
"description": "endPublishedDate"
},
"moderation": {
"type": "boolean",
"description": "moderation"
},
"contents": {
"type": "object",
"description": "Content enrichment options. Same shape as in /search and /contents."
}
},
"required": [
"url"
]
}
},
{
"name": "superagnt_connection_exa_answer",
"description": "Generate an answer from search results",
"method": "POST",
"path": "/answer",
"parameters": {
"type": "object",
"properties": {
"query": {
"type": "string",
"description": "The question or query to answer."
},
"stream": {
"type": "boolean",
"description": "If true, returns an SSE stream."
},
"text": {
"type": "boolean",
"description": "If true, citations include full text content of each source."
},
"outputSchema": {
"type": "object",
"description": "JSON Schema Draft 7 for structured answers."
}
},
"required": [
"query"
]
}
}
]
Example
curl -X POST 'https://api.superagnt.com/v1/connections/exa/search' \
-H 'Authorization: Bearer $SUPERAGNT_API_KEY' \
-H 'Content-Type: application/json' \
-d '{"query":"latest developments in LLMs","numResults":10}'
Use Cases
- Give agents real-time web knowledge beyond their training cutoff
- Research a company, competitor, or topic with cited sources and summaries
- Build retrieval-augmented generation pipelines grounded in the live web
- Pull clean page text for summarization, extraction, or ingest into a vector store
- Monitor news and blog posts filtered to trusted domains and date ranges
- Run agentic Q&A that returns answers plus the sources Exa used
Links
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 469k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 321k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 321k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 321k | Updated | Claude CodeCodex | Verified | Install |