[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6MD3Q49DXZZ567EN0CGPYVK":3,"skill-same:01M6MD3Q49DXZZ567EN0CGPYVK":69},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":41,"linkOnly":42,"markdownContent":43,"markdownBody":44,"frontMatter":45,"refName":46,"commitSha":47,"skillDirectory":48,"forksCount":49,"repositoryHtmlUrl":50,"repositoryHomepageUrl":51,"ownerHtmlUrl":52,"ownerBlogUrl":51,"ownerDescription":53,"updatedAt":54,"install":55,"resources":68},"01M6MD3Q49DXZZ567EN0CGPYVK","exa-search",null,"Call Exa Search directly with cURL or raw HTTP. Use when an agent needs Exa semantic web retrieval from POST \u002Fsearch without an SDK, including ranked results, domain or category filters, freshness-aware result content, highlights or text extraction, structured output, or streaming search responses.","compozy\u002Fcompozy","compozy","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F205107042?v=4","01M6JWVTXX4DX38AWZ7SPFEKEV","CompozyOS","01M6JWVV2P34GM1PDATVB2EV05",2871,"MIT","https:\u002F\u002Fgithub.com\u002Fcompozy\u002Fcompozy\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fexa-search\u002FSKILL.md",[],[19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40],"search","exa search","exa","filters","when","result","with","output","retrieval","post","structured output","domain","call","sdk","use","results","curl","text","content","streaming","http","web","2026-11-10 10:12:30",false,"---\nname: exa-search\ndescription: \"Call Exa Search directly with cURL or raw HTTP. Use when an agent needs Exa semantic web retrieval from POST \u002Fsearch without an SDK, including ranked results, domain or category filters, freshness-aware result content, highlights or text extraction, structured output, or streaming search responses.\"\n---\n\n# Exa Search\n\n> Requires API key: Get one at https:\u002F\u002Fdashboard.exa.ai\u002Fapi-keys\n>\n> Header: `x-api-key: $EXA_API_KEY`\n\nUse `POST https:\u002F\u002Fapi.exa.ai\u002Fsearch` for semantic web retrieval, ranked results, and optional result-level extraction in one raw HTTP call. Start with `type: \"auto\"` for general retrieval. Add `contents` only when the caller needs page text, highlights, summaries, freshness-controlled crawling, subpages, or extracted links.\n\n## Quick Start (cURL)\n\n### Basic search\n\n```bash\ncurl -sS -X POST \"https:\u002F\u002Fapi.exa.ai\u002Fsearch\" \\\n  -H \"Content-Type: application\u002Fjson\" \\\n  -H \"x-api-key: $EXA_API_KEY\" \\\n  -d '{\n    \"query\": \"latest developments in LLMs\",\n    \"type\": \"auto\",\n    \"numResults\": 10\n  }'\n```\n\n### Search with highlights\n\n```bash\ncurl -sS -X POST \"https:\u002F\u002Fapi.exa.ai\u002Fsearch\" \\\n  -H \"Content-Type: application\u002Fjson\" \\\n  -H \"x-api-key: $EXA_API_KEY\" \\\n  -d '{\n    \"query\": \"latest developments in LLMs\",\n    \"type\": \"auto\",\n    \"numResults\": 5,\n    \"contents\": {\n      \"highlights\": true\n    }\n  }'\n```\n\n### With filters and freshness\n\n```bash\ncurl -sS -X POST \"https:\u002F\u002Fapi.exa.ai\u002Fsearch\" \\\n  -H \"Content-Type: application\u002Fjson\" \\\n  -H \"x-api-key: $EXA_API_KEY\" \\\n  -d '{\n    \"query\": \"AI regulation policy updates\",\n    \"type\": \"auto\",\n    \"category\": \"news\",\n    \"numResults\": 10,\n    \"includeDomains\": [\"reuters.com\", \"bbc.com\"],\n    \"startPublishedDate\": \"2025-01-01\",\n    \"contents\": {\n      \"text\": {\n        \"maxCharacters\": 2000\n      },\n      \"maxAgeHours\": 24,\n      \"livecrawlTimeout\": 12000\n    }\n  }'\n```\n\n### Deep search\n\n```bash\ncurl -sS -X POST \"https:\u002F\u002Fapi.exa.ai\u002Fsearch\" \\\n  -H \"Content-Type: application\u002Fjson\" \\\n  -H \"x-api-key: $EXA_API_KEY\" \\\n  -d '{\n    \"query\": \"map the major technical and commercial tradeoffs in sodium-ion batteries for grid storage\",\n    \"type\": \"deep\",\n    \"numResults\": 8\n  }'\n```\n\n## Endpoint\n\n```text\nPOST https:\u002F\u002Fapi.exa.ai\u002Fsearch\n```\n\nAuthentication: `x-api-key: \u003CAPI_KEY>` header. Exa also accepts `Authorization: Bearer \u003CAPI_KEY>`, but prefer `x-api-key` in cURL examples for consistency.\n\nUse this endpoint when the agent needs search results. If the agent already has URLs and only needs extraction, use `POST \u002Fcontents` instead.\n\n## Parameters\n\n### Core request parameters\n\n| Parameter | Type | Required | Default | Description |\n| --- | --- | --- | --- | --- |\n| `query` | string | Yes | - | Natural-language search query. Long, semantically rich descriptions work well. |\n| `type` | string | No | `auto` | Search method: `auto`, `fast`, `instant`, `deep-lite`, `deep`, or `deep-reasoning`. |\n| `numResults` | integer | No | `10` | Number of results to return. Use small values for agent loops; maximum is 100. |\n| `category` | string | No | - | Specialized result type: `company`, `people`, `research paper`, `news`, `personal site`, or `financial report`. |\n| `includeDomains` | string[] | No | - | Only return results from these domains, paths, or wildcard patterns. Max 1200. |\n| `excludeDomains` | string[] | No | - | Exclude these domains, paths, or wildcard patterns. Max 1200. |\n| `startPublishedDate` | string | No | - | ISO 8601 lower bound for result publication date. |\n| `endPublishedDate` | string | No | - | ISO 8601 upper bound for result publication date. |\n| `userLocation` | string | No | - | Two-letter ISO country code such as `US` or `GB`. |\n| `moderation` | boolean | No | `false` | Filter unsafe content from results. |\n| `additionalQueries` | string[] | No | - | Extra query variants for deep-search variants. Use alongside the main `query`. |\n| `systemPrompt` | string | No | - | Instructions for synthesized output and deep-search planning, such as source preferences. |\n| `outputSchema` | object | No | - | JSON Schema controlling `output.content`. Adds synthesized output and grounding. |\n| `stream` | boolean | No | `false` | If `true`, returns SSE instead of a single JSON response. |\n| `compliance` | string | No | - | Enterprise-only compliance mode, such as `hipaa`, when enabled for the account. |\n\n### Content parameters nested under `contents`\n\nOn `\u002Fsearch`, `text`, `highlights`, and `summary` must be nested under `contents`.\n\n| Parameter | Type | Required | Default | Description |\n| --- | --- | --- | --- | --- |\n| `contents.text` | boolean or object | No | - | Return full page text as markdown. Object form supports `maxCharacters`, `includeHtmlTags`, `verbosity`, `includeSections`, and `excludeSections`. |\n| `contents.highlights` | boolean or object | No | - | Return query-relevant excerpts. Prefer `true` for agent workflows unless a fixed character budget is required. |\n| `contents.summary` | boolean or object | No | - | Return per-result LLM summaries. Use sparingly because each result adds synthesis work. |\n| `contents.maxAgeHours` | integer | No | - | Freshness control. `0` always live crawls; `-1` uses cache only; omit for default cache-first behavior with crawl fallback. |\n| `contents.livecrawlTimeout` | integer | No | `10000` | Timeout for live crawling in milliseconds. Use `10000` to `15000` for most freshness-sensitive calls. |\n| `contents.subpages` | integer | No | `0` | Number of linked subpages to crawl per result. |\n| `contents.subpageTarget` | string or string[] | No | - | Terms used to prioritize which subpages matter, such as `[\"api\", \"pricing\"]`. |\n| `contents.extras.links` | integer | No | `0` | Number of links to extract from each result page. |\n| `contents.extras.imageLinks` | integer | No | `0` | Number of image URLs to extract from each result page. |\n\n### Text object options\n\n| Parameter | Type | Default | Description |\n| --- | --- | --- | --- |\n| `maxCharacters` | integer | - | Character limit for returned text. Use this instead of `tokensNum`. |\n| `includeHtmlTags` | boolean | `false` | Preserve HTML tags in output. |\n| `verbosity` | string | `compact` | `compact`, `standard`, or `full`. Pair fresh section-aware extraction with `contents.maxAgeHours: 0`. |\n| `includeSections` | string[] | - | Only include selected sections: `header`, `navigation`, `banner`, `body`, `sidebar`, `footer`, `metadata`. |\n| `excludeSections` | string[] | - | Exclude selected sections from the same section list. |\n\n### Highlights object options\n\nPrefer `contents.highlights: true` for the highest-quality default. Only use object form when the agent needs a custom focus or budget.\n\n| Parameter | Type | Default | Description |\n| --- | --- | --- | --- |\n| `query` | string | - | Custom query guiding which excerpts are returned. |\n| `maxCharacters` | integer | - | Cap highlight characters per URL. Omit unless the caller has a strict budget. |\n\n### Summary object options\n\n| Parameter | Type | Default | Description |\n| --- | --- | --- | --- |\n| `query` | string | - | Custom query for the summary. |\n| `schema` | object | - | JSON Schema for structured per-result summaries. |\n\n## Search Types\n\nSearch `type` controls the retrieval and synthesis mode. Pick the mode for the workflow, not just the output format. `outputSchema` can be used with any search type; use deeper modes when the search process itself needs more planning, synthesis, or reasoning.\n\n| Type | Best for | Tradeoff |\n| --- | --- | --- |\n| `auto` | General default search and most new integrations | Balances speed and quality without requiring the caller to tune retrieval strategy. |\n| `fast` | Low-latency agent loops and product paths | Faster than `auto`; use when responsiveness matters more than maximum reasoning depth. |\n| `instant` | Real-time UI, chat, voice, and autocomplete-style paths | Lowest latency path; use for quick retrieval rather than deep synthesis. |\n| `deep-lite` | Lightweight research or synthesis | Adds more planning and synthesis than `auto` while staying lighter than full `deep`. |\n| `deep` | Multi-step research, comparisons, and synthesis-heavy retrieval | Higher latency; better when the query needs exploration across several sources. |\n| `deep-reasoning` | Hard research tasks with high ambiguity or complex tradeoffs | Highest latency and reasoning depth. |\n\nUse `auto` unless latency or reasoning depth is the primary constraint. Use `fast` or `instant` for time-sensitive calls. Use `deep`, `deep-lite`, or `deep-reasoning` when the query needs multi-step source discovery, comparison, or synthesis.\n\n### Mode-only examples\n\n```json\n{\n  \"query\": \"recent product launches from major AI chip companies\",\n  \"type\": \"fast\",\n  \"numResults\": 5\n}\n```\n\n```json\n{\n  \"query\": \"compare competing explanations for the recent rise in grid-scale battery deployments\",\n  \"type\": \"deep\",\n  \"numResults\": 8\n}\n```\n\n## Structured Output\n\nUse `systemPrompt` for behavior and `outputSchema` for shape.\n\n```bash\ncurl -sS -X POST \"https:\u002F\u002Fapi.exa.ai\u002Fsearch\" \\\n  -H \"Content-Type: application\u002Fjson\" \\\n  -H \"x-api-key: $EXA_API_KEY\" \\\n  -d '{\n    \"query\": \"compare the latest frontier AI model releases\",\n    \"type\": \"deep\",\n    \"systemPrompt\": \"Prefer official sources and avoid duplicate results.\",\n    \"outputSchema\": {\n      \"type\": \"object\",\n      \"properties\": {\n        \"models\": {\n          \"type\": \"array\",\n          \"items\": {\n            \"type\": \"object\",\n            \"properties\": {\n              \"name\": { \"type\": \"string\" },\n              \"notable_claims\": {\n                \"type\": \"array\",\n                \"items\": { \"type\": \"string\" }\n              }\n            },\n            \"required\": [\"name\", \"notable_claims\"]\n          }\n        }\n      },\n      \"required\": [\"models\"]\n    },\n    \"contents\": {\n      \"highlights\": true\n    }\n  }'\n```\n\nKeep schemas compact and bounded. Do not add citation fields to the schema; grounding is returned separately in `output.grounding`.\n\n## Streaming\n\nStreaming applies to synthesized output, so include `outputSchema` along with `-N`, `Accept: text\u002Fevent-stream`, and `stream: true`. Without `outputSchema`, the endpoint returns the normal JSON search response even when `stream` is `true`.\n\n```bash\ncurl -sS -N -X POST \"https:\u002F\u002Fapi.exa.ai\u002Fsearch\" \\\n  -H \"Content-Type: application\u002Fjson\" \\\n  -H \"Accept: text\u002Fevent-stream\" \\\n  -H \"x-api-key: $EXA_API_KEY\" \\\n  -d '{\n    \"query\": \"recent grid-scale battery deployments\",\n    \"type\": \"deep\",\n    \"stream\": true,\n    \"outputSchema\": {\n      \"type\": \"object\",\n      \"properties\": {\n        \"summary\": { \"type\": \"string\" }\n      },\n      \"required\": [\"summary\"]\n    },\n    \"contents\": {\n      \"highlights\": true\n    }\n  }'\n```\n\nTreat streaming as SSE rather than JSON. Each `data:` frame contains an OpenAI-compatible chat completion chunk; read partial text from `choices[0].delta.content` and handle completion or error frames defensively.\n\n## Response Fields\n\n| Field | Type | Description |\n| --- | --- | --- |\n| `requestId` | string | Unique request identifier. |\n| `results` | array | Ranked result objects. |\n| `results[].title` | string | Page title. |\n| `results[].url` | string | Page URL. |\n| `results[].publishedDate` | string or null | Estimated publication date when available. |\n| `results[].author` | string or null | Author when available. |\n| `results[].text` | string | Returned when `contents.text` is requested. |\n| `results[].highlights` | string[] | Returned when `contents.highlights` is requested. |\n| `results[].highlightScores` | number[] | Similarity scores for highlights. |\n| `results[].summary` | string | Returned when `contents.summary` is requested. |\n| `results[].subpages` | array | Nested result objects from subpage crawling. |\n| `results[].extras.links` | string[] | Extracted links when requested. |\n| `output.content` | string or object | Synthesized output when `outputSchema` is provided. |\n| `output.grounding` | array | Citations and confidence labels for synthesized fields. |\n| `costDollars.total` | number | Total request cost when returned. |\n| `searchTime` | number | Search latency when returned. |\n\n## Critical Pitfalls\n\n- Keep `text`, `highlights`, and `summary` inside `contents` on `\u002Fsearch`.\n- Do not send top-level `text`, `highlights`, or `summary`; that shape belongs to `\u002Fcontents`.\n- Do not send `tokensNum`; use `contents.text.maxCharacters` to cap extracted text.\n- Do not use `useAutoprompt`, `numSentences`, or `highlightsPerUrl` in new requests.\n- Use `contents.maxAgeHours` instead of `livecrawl`.\n- Use documented categories only: `company`, `people`, `research paper`, `news`, `personal site`, and `financial report`.\n- Avoid invalid category\u002Ffilter combinations. `company` and `people` do not support `startPublishedDate` or `endPublishedDate`. `company` supports `excludeDomains`; `people` does not, and `people` only accepts LinkedIn domains in `includeDomains`.\n- Pick one of `contents.highlights`, `contents.text`, or `contents.summary` by default. Stack modes only when the caller truly needs multiple views of each page.\n- Expect SSE only when `stream: true` is paired with `outputSchema`; otherwise `\u002Fsearch` returns its normal JSON response.\n","# Exa Search\n\n> Requires API key: Get one at https:\u002F\u002Fdashboard.exa.ai\u002Fapi-keys\n>\n> Header: `x-api-key: $EXA_API_KEY`\n\nUse `POST https:\u002F\u002Fapi.exa.ai\u002Fsearch` for semantic web retrieval, ranked results, and optional result-level extraction in one raw HTTP call. Start with `type: \"auto\"` for general retrieval. Add `contents` only when the caller needs page text, highlights, summaries, freshness-controlled crawling, subpages, or extracted links.\n\n## Quick Start (cURL)\n\n### Basic search\n\n```bash\ncurl -sS -X POST \"https:\u002F\u002Fapi.exa.ai\u002Fsearch\" \\\n  -H \"Content-Type: application\u002Fjson\" \\\n  -H \"x-api-key: $EXA_API_KEY\" \\\n  -d '{\n    \"query\": \"latest developments in LLMs\",\n    \"type\": \"auto\",\n    \"numResults\": 10\n  }'\n```\n\n### Search with highlights\n\n```bash\ncurl -sS -X POST \"https:\u002F\u002Fapi.exa.ai\u002Fsearch\" \\\n  -H \"Content-Type: application\u002Fjson\" \\\n  -H \"x-api-key: $EXA_API_KEY\" \\\n  -d '{\n    \"query\": \"latest developments in LLMs\",\n    \"type\": \"auto\",\n    \"numResults\": 5,\n    \"contents\": {\n      \"highlights\": true\n    }\n  }'\n```\n\n### With filters and freshness\n\n```bash\ncurl -sS -X POST \"https:\u002F\u002Fapi.exa.ai\u002Fsearch\" \\\n  -H \"Content-Type: application\u002Fjson\" \\\n  -H \"x-api-key: $EXA_API_KEY\" \\\n  -d '{\n    \"query\": \"AI regulation policy updates\",\n    \"type\": \"auto\",\n    \"category\": \"news\",\n    \"numResults\": 10,\n    \"includeDomains\": [\"reuters.com\", \"bbc.com\"],\n    \"startPublishedDate\": \"2025-01-01\",\n    \"contents\": {\n      \"text\": {\n        \"maxCharacters\": 2000\n      },\n      \"maxAgeHours\": 24,\n      \"livecrawlTimeout\": 12000\n    }\n  }'\n```\n\n### Deep search\n\n```bash\ncurl -sS -X POST \"https:\u002F\u002Fapi.exa.ai\u002Fsearch\" \\\n  -H \"Content-Type: application\u002Fjson\" \\\n  -H \"x-api-key: $EXA_API_KEY\" \\\n  -d '{\n    \"query\": \"map the major technical and commercial tradeoffs in sodium-ion batteries for grid storage\",\n    \"type\": \"deep\",\n    \"numResults\": 8\n  }'\n```\n\n## Endpoint\n\n```text\nPOST https:\u002F\u002Fapi.exa.ai\u002Fsearch\n```\n\nAuthentication: `x-api-key: \u003CAPI_KEY>` header. Exa also accepts `Authorization: Bearer \u003CAPI_KEY>`, but prefer `x-api-key` in cURL examples for consistency.\n\nUse this endpoint when the agent needs search results. If the agent already has URLs and only needs extraction, use `POST \u002Fcontents` instead.\n\n## Parameters\n\n### Core request parameters\n\n| Parameter | Type | Required | Default | Description |\n| --- | --- | --- | --- | --- |\n| `query` | string | Yes | - | Natural-language search query. Long, semantically rich descriptions work well. |\n| `type` | string | No | `auto` | Search method: `auto`, `fast`, `instant`, `deep-lite`, `deep`, or `deep-reasoning`. |\n| `numResults` | integer | No | `10` | Number of results to return. Use small values for agent loops; maximum is 100. |\n| `category` | string | No | - | Specialized result type: `company`, `people`, `research paper`, `news`, `personal site`, or `financial report`. |\n| `includeDomains` | string[] | No | - | Only return results from these domains, paths, or wildcard patterns. Max 1200. |\n| `excludeDomains` | string[] | No | - | Exclude these domains, paths, or wildcard patterns. Max 1200. |\n| `startPublishedDate` | string | No | - | ISO 8601 lower bound for result publication date. |\n| `endPublishedDate` | string | No | - | ISO 8601 upper bound for result publication date. |\n| `userLocation` | string | No | - | Two-letter ISO country code such as `US` or `GB`. |\n| `moderation` | boolean | No | `false` | Filter unsafe content from results. |\n| `additionalQueries` | string[] | No | - | Extra query variants for deep-search variants. Use alongside the main `query`. |\n| `systemPrompt` | string | No | - | Instructions for synthesized output and deep-search planning, such as source preferences. |\n| `outputSchema` | object | No | - | JSON Schema controlling `output.content`. Adds synthesized output and grounding. |\n| `stream` | boolean | No | `false` | If `true`, returns SSE instead of a single JSON response. |\n| `compliance` | string | No | - | Enterprise-only compliance mode, such as `hipaa`, when enabled for the account. |\n\n### Content parameters nested under `contents`\n\nOn `\u002Fsearch`, `text`, `highlights`, and `summary` must be nested under `contents`.\n\n| Parameter | Type | Required | Default | Description |\n| --- | --- | --- | --- | --- |\n| `contents.text` | boolean or object | No | - | Return full page text as markdown. Object form supports `maxCharacters`, `includeHtmlTags`, `verbosity`, `includeSections`, and `excludeSections`. |\n| `contents.highlights` | boolean or object | No | - | Return query-relevant excerpts. Prefer `true` for agent workflows unless a fixed character budget is required. |\n| `contents.summary` | boolean or object | No | - | Return per-result LLM summaries. Use sparingly because each result adds synthesis work. |\n| `contents.maxAgeHours` | integer | No | - | Freshness control. `0` always live crawls; `-1` uses cache only; omit for default cache-first behavior with crawl fallback. |\n| `contents.livecrawlTimeout` | integer | No | `10000` | Timeout for live crawling in milliseconds. Use `10000` to `15000` for most freshness-sensitive calls. |\n| `contents.subpages` | integer | No | `0` | Number of linked subpages to crawl per result. |\n| `contents.subpageTarget` | string or string[] | No | - | Terms used to prioritize which subpages matter, such as `[\"api\", \"pricing\"]`. |\n| `contents.extras.links` | integer | No | `0` | Number of links to extract from each result page. |\n| `contents.extras.imageLinks` | integer | No | `0` | Number of image URLs to extract from each result page. |\n\n### Text object options\n\n| Parameter | Type | Default | Description |\n| --- | --- | --- | --- |\n| `maxCharacters` | integer | - | Character limit for returned text. Use this instead of `tokensNum`. |\n| `includeHtmlTags` | boolean | `false` | Preserve HTML tags in output. |\n| `verbosity` | string | `compact` | `compact`, `standard`, or `full`. Pair fresh section-aware extraction with `contents.maxAgeHours: 0`. |\n| `includeSections` | string[] | - | Only include selected sections: `header`, `navigation`, `banner`, `body`, `sidebar`, `footer`, `metadata`. |\n| `excludeSections` | string[] | - | Exclude selected sections from the same section list. |\n\n### Highlights object options\n\nPrefer `contents.highlights: true` for the highest-quality default. Only use object form when the agent needs a custom focus or budget.\n\n| Parameter | Type | Default | Description |\n| --- | --- | --- | --- |\n| `query` | string | - | Custom query guiding which excerpts are returned. |\n| `maxCharacters` | integer | - | Cap highlight characters per URL. Omit unless the caller has a strict budget. |\n\n### Summary object options\n\n| Parameter | Type | Default | Description |\n| --- | --- | --- | --- |\n| `query` | string | - | Custom query for the summary. |\n| `schema` | object | - | JSON Schema for structured per-result summaries. |\n\n## Search Types\n\nSearch `type` controls the retrieval and synthesis mode. Pick the mode for the workflow, not just the output format. `outputSchema` can be used with any search type; use deeper modes when the search process itself needs more planning, synthesis, or reasoning.\n\n| Type | Best for | Tradeoff |\n| --- | --- | --- |\n| `auto` | General default search and most new integrations | Balances speed and quality without requiring the caller to tune retrieval strategy. |\n| `fast` | Low-latency agent loops and product paths | Faster than `auto`; use when responsiveness matters more than maximum reasoning depth. |\n| `instant` | Real-time UI, chat, voice, and autocomplete-style paths | Lowest latency path; use for quick retrieval rather than deep synthesis. |\n| `deep-lite` | Lightweight research or synthesis | Adds more planning and synthesis than `auto` while staying lighter than full `deep`. |\n| `deep` | Multi-step research, comparisons, and synthesis-heavy retrieval | Higher latency; better when the query needs exploration across several sources. |\n| `deep-reasoning` | Hard research tasks with high ambiguity or complex tradeoffs | Highest latency and reasoning depth. |\n\nUse `auto` unless latency or reasoning depth is the primary constraint. Use `fast` or `instant` for time-sensitive calls. Use `deep`, `deep-lite`, or `deep-reasoning` when the query needs multi-step source discovery, comparison, or synthesis.\n\n### Mode-only examples\n\n```json\n{\n  \"query\": \"recent product launches from major AI chip companies\",\n  \"type\": \"fast\",\n  \"numResults\": 5\n}\n```\n\n```json\n{\n  \"query\": \"compare competing explanations for the recent rise in grid-scale battery deployments\",\n  \"type\": \"deep\",\n  \"numResults\": 8\n}\n```\n\n## Structured Output\n\nUse `systemPrompt` for behavior and `outputSchema` for shape.\n\n```bash\ncurl -sS -X POST \"https:\u002F\u002Fapi.exa.ai\u002Fsearch\" \\\n  -H \"Content-Type: application\u002Fjson\" \\\n  -H \"x-api-key: $EXA_API_KEY\" \\\n  -d '{\n    \"query\": \"compare the latest frontier AI model releases\",\n    \"type\": \"deep\",\n    \"systemPrompt\": \"Prefer official sources and avoid duplicate results.\",\n    \"outputSchema\": {\n      \"type\": \"object\",\n      \"properties\": {\n        \"models\": {\n          \"type\": \"array\",\n          \"items\": {\n            \"type\": \"object\",\n            \"properties\": {\n              \"name\": { \"type\": \"string\" },\n              \"notable_claims\": {\n                \"type\": \"array\",\n                \"items\": { \"type\": \"string\" }\n              }\n            },\n            \"required\": [\"name\", \"notable_claims\"]\n          }\n        }\n      },\n      \"required\": [\"models\"]\n    },\n    \"contents\": {\n      \"highlights\": true\n    }\n  }'\n```\n\nKeep schemas compact and bounded. Do not add citation fields to the schema; grounding is returned separately in `output.grounding`.\n\n## Streaming\n\nStreaming applies to synthesized output, so include `outputSchema` along with `-N`, `Accept: text\u002Fevent-stream`, and `stream: true`. Without `outputSchema`, the endpoint returns the normal JSON search response even when `stream` is `true`.\n\n```bash\ncurl -sS -N -X POST \"https:\u002F\u002Fapi.exa.ai\u002Fsearch\" \\\n  -H \"Content-Type: application\u002Fjson\" \\\n  -H \"Accept: text\u002Fevent-stream\" \\\n  -H \"x-api-key: $EXA_API_KEY\" \\\n  -d '{\n    \"query\": \"recent grid-scale battery deployments\",\n    \"type\": \"deep\",\n    \"stream\": true,\n    \"outputSchema\": {\n      \"type\": \"object\",\n      \"properties\": {\n        \"summary\": { \"type\": \"string\" }\n      },\n      \"required\": [\"summary\"]\n    },\n    \"contents\": {\n      \"highlights\": true\n    }\n  }'\n```\n\nTreat streaming as SSE rather than JSON. Each `data:` frame contains an OpenAI-compatible chat completion chunk; read partial text from `choices[0].delta.content` and handle completion or error frames defensively.\n\n## Response Fields\n\n| Field | Type | Description |\n| --- | --- | --- |\n| `requestId` | string | Unique request identifier. |\n| `results` | array | Ranked result objects. |\n| `results[].title` | string | Page title. |\n| `results[].url` | string | Page URL. |\n| `results[].publishedDate` | string or null | Estimated publication date when available. |\n| `results[].author` | string or null | Author when available. |\n| `results[].text` | string | Returned when `contents.text` is requested. |\n| `results[].highlights` | string[] | Returned when `contents.highlights` is requested. |\n| `results[].highlightScores` | number[] | Similarity scores for highlights. |\n| `results[].summary` | string | Returned when `contents.summary` is requested. |\n| `results[].subpages` | array | Nested result objects from subpage crawling. |\n| `results[].extras.links` | string[] | Extracted links when requested. |\n| `output.content` | string or object | Synthesized output when `outputSchema` is provided. |\n| `output.grounding` | array | Citations and confidence labels for synthesized fields. |\n| `costDollars.total` | number | Total request cost when returned. |\n| `searchTime` | number | Search latency when returned. |\n\n## Critical Pitfalls\n\n- Keep `text`, `highlights`, and `summary` inside `contents` on `\u002Fsearch`.\n- Do not send top-level `text`, `highlights`, or `summary`; that shape belongs to `\u002Fcontents`.\n- Do not send `tokensNum`; use `contents.text.maxCharacters` to cap extracted text.\n- Do not use `useAutoprompt`, `numSentences`, or `highlightsPerUrl` in new requests.\n- Use `contents.maxAgeHours` instead of `livecrawl`.\n- Use documented categories only: `company`, `people`, `research paper`, `news`, `personal site`, and `financial report`.\n- Avoid invalid category\u002Ffilter combinations. `company` and `people` do not support `startPublishedDate` or `endPublishedDate`. `company` supports `excludeDomains`; `people` does not, and `people` only accepts LinkedIn domains in `includeDomains`.\n- Pick one of `contents.highlights`, `contents.text`, or `contents.summary` by default. Stack modes only when the caller truly needs multiple views of each page.\n- Expect SSE only when `stream: true` is paired with `outputSchema`; otherwise `\u002Fsearch` returns its normal JSON response.\n",{"name":5,"description":7},"main","65613632bb94072e27fc5015d17387e28a870111",".agents\u002Fskills\u002Fexa-search",187,"https:\u002F\u002Fgithub.com\u002Fcompozy\u002Fcompozy","https:\u002F\u002Fcompozy.com","https:\u002F\u002Fgithub.com\u002Fcompozy","An OS for AI agents","2026-11-15 03:55:11",{"source":8,"skillName":5,"commands":56,"prompt":66,"zipAvailable":67},[57,60,63],{"tool":58,"command":59},"npx","npx skills add compozy\u002Fcompozy --skill exa-search",{"tool":61,"command":62},"bunx","bunx skills add compozy\u002Fcompozy --skill exa-search",{"tool":64,"command":65},"pnpm","pnpm dlx skills add compozy\u002Fcompozy --skill exa-search","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fcompozy\u002Fcompozy\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fexa-search\u002FSKILL.md",true,[],[70,104,129,151,162,174],{"skillNo":71,"name":72,"nameZh":6,"description":73,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":74,"licenseFileUrl":75,"sourceHtmlUrl":76,"categories":77,"keywords":83,"publishedAt":41},"01M6MD7W312X49RNEJRKKVKTTD","security-review","Security code review for vulnerabilities. Use when asked to \"security review\", \"find vulnerabilities\", \"check for security issues\", \"audit security\", \"OWASP review\", or review code for injection, XSS, authentication, authorization, cryptography issues. Provides systematic review with confidence-based reporting.","CC-BY-SA-4.0","https:\u002F\u002Fgithub.com\u002Fcompozy\u002Fcompozy\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fsecurity-review\u002FLICENSE","https:\u002F\u002Fgithub.com\u002Fcompozy\u002Fcompozy\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fsecurity-review\u002FSKILL.md",[78],{"categoryNo":79,"categoryName":80,"taxonomyType":81,"taxonomyTypeDesc":82},"01JBSEEDCAT000000000000007","代码审查","SCENARIO","使用场景",[84,85,86,87,88,89,90,91,92,93,94,95,96,97,98,99,100,101,33,102,23,103],"security","security review","review","injection","authentication","code review","issues","Bash","code","xss","audit","systematic","read","reporting","cryptography","owasp","systematic review","for","task","audit security",{"skillNo":105,"name":106,"nameZh":6,"description":107,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":108,"categories":109,"keywords":115,"publishedAt":41},"01M6MD5K93G1GP8VA56V2FEPZV","lesson-learned","Extract evidence-backed engineering lessons from recent changes or incidents. Use for requested reflection, lesson updates, or retiring an obsolete workflow rule.","https:\u002F\u002Fgithub.com\u002Fcompozy\u002Fcompozy\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Flesson-learned\u002FSKILL.md",[110],{"categoryNo":111,"categoryName":112,"taxonomyType":113,"taxonomyTypeDesc":114},"01JBSEEDCAT000000000000039","教师与教育","PROFESSION","适用职业",[116,117,118,101,119,120,121,122,123,124,125,126,33,127,128],"learned","lesson learned","lesson","changes","evidence","incidents","workflow","engineering","updates","extract","rule","lessons","reflection",{"skillNo":130,"name":9,"nameZh":6,"description":131,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":132,"categories":133,"keywords":134,"publishedAt":41},"01M6MDPNH3J60VRCGTPHVTHZR8","Operate CompozyOS. Use when working with CompozyOS sessions, agents, native tools, skills, tasks, Loops, Goals, Terminal, desktops and windows, automation, extensions, or configuration. Don't use for unrelated projects.","https:\u002F\u002Fgithub.com\u002Fcompozy\u002Fcompozy\u002Fblob\u002Fmain\u002Fskills\u002Fcompozy\u002FSKILL.md",[],[9,135,136,137,138,101,139,140,33,141,142,143,25,144,145,146,147,148,149,150,23],"sessions","windows","projects","operate","runtime","true","automation","tasks","extensions","loops","unrelated","goals","native","terminal","configuration","Tools",{"skillNo":152,"name":153,"nameZh":6,"description":154,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":155,"categories":156,"keywords":157,"publishedAt":41},"01M6MDPHKNPQHJ1TF21E1HSFNF","remotion-upgrade","Upgrade Remotion, and related packages","https:\u002F\u002Fgithub.com\u002Fcompozy\u002Fcompozy\u002Fblob\u002Fmain\u002Fpackages\u002Flaunch-video\u002F.agents\u002Fskills\u002Fremotion-upgrade\u002FSKILL.md",[],[158,159,160,161],"remotion upgrade","upgrade","Remotion","packages",{"skillNo":163,"name":164,"nameZh":6,"description":165,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":166,"categories":167,"keywords":168,"publishedAt":41},"01M6MDPFT91ZWC7Z09MRNZPAJG","remotion-studio","Preview a Remotion video","https:\u002F\u002Fgithub.com\u002Fcompozy\u002Fcompozy\u002Fblob\u002Fmain\u002Fpackages\u002Flaunch-video\u002F.agents\u002Fskills\u002Fremotion-studio\u002FSKILL.md",[],[169,170,160,171,172,173],"studio","remotion studio","Video","preview","remotion video",{"skillNo":175,"name":176,"nameZh":6,"description":177,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":178,"categories":179,"keywords":180,"publishedAt":41},"01M6MDPD5XDZPYB5NAZGKMPXWC","remotion-saas","Build an app with Remotion","https:\u002F\u002Fgithub.com\u002Fcompozy\u002Fcompozy\u002Fblob\u002Fmain\u002Fpackages\u002Flaunch-video\u002F.agents\u002Fskills\u002Fremotion-saas\u002FSKILL.md",[],[181,182,160,25,183,184],"saas","remotion saas","app","build"]