Built by Metorial, the integration platform for agentic AI.
Remove an enrichment from a Webset. Cancels a running enrichment and removes its existing enrichment results; this does not undo consumed credits or history.
Retrieve a specific item from a Webset by its ID, including all properties and enrichment results.
Find web pages similar in meaning to a given URL. Useful for competitor analysis, discovering related content, or building recommendation systems. Supports filtering by domains, dates, and text content. Optionally retrieve full text, highlights, and summaries. Exa documents this endpoint as deprecated while retaining it; prefer Web Search with a query describing the source for new workflows.
List all Websets in your account. Supports cursor-based pagination.
Read historical Webset export metadata. This compatibility route is unverified in current public documentation; no file availability or renewal is promised. A returned URL is legacy metadata.
Create a new Webset — a structured collection of web entities discovered by Exa's search agents. Define a search query and criteria; Exa finds matching web entities and verifies them. Optionally include enrichments to extract additional structured data from found items.
Request the historical Webset export route for CSV, JSON or XLSX. This compatibility route is not corroborated by the current public API or SDK; confirm availability with Exa before use. The export runs asynchronously — use the returned export ID to check completion status.
Start an asynchronous in-depth web research task. Exa explores the web, gathers sources, and synthesizes findings into a structured report with citations. Creates a retained, billable task. A failed receipt does not prove the task was not created; inspect before retrying. Returns a research ID for polling the status. Use the **Get Research Status** tool to check completion and retrieve results. Optionally provide an output schema for structured JSON results.
Update a Webset's title or metadata. externalId is retained as a compatibility field but the current update endpoint does not support changing it.
Add an enrichment to a Webset. Enrichments are agents that search the web to enhance Webset items with additional structured data. Define what data to extract and its expected format (text, number, date, etc.). The enrichment runs on all existing and future items.
List items in a Webset. Each item represents a web entity found and verified by Exa's search agents. Items include source URLs, properties, and enrichment results. Supports cursor-based pagination and native sourceId filtering.
Cancel all running operations on a Webset, including active searches and enrichments.
Permanently delete a Webset and all its items, searches, enrichments, and monitors.
Delete a monitor from a Webset, stopping its scheduled operations.
Create a monitor on a Webset to automatically keep it updated on a schedule. **Search** behavior runs recurring, billable searches with automatic deduplication. The retained refresh input is unsupported by the current native endpoint and is refused before creation. Schedules use cron expressions.
Retrieve content from a list of URLs without performing a search. Extracts full text as markdown, key highlights, and AI-generated summaries from web pages. Useful when you already have specific URLs and need to extract their content.
Check the status of an asynchronous research task and retrieve its results when completed. Returns the current status, and when completed, the full research output with content and cost information.
Remove a specific item from a Webset.
Retrieve details of a specific Webset by its ID, including status, metadata, and associated searches, enrichments, and monitors.
Get a direct AI-generated answer to a question with citations from the web. Exa searches the web and synthesizes an answer backed by source references. Ideal for factual questions, research queries, and getting quick answers grounded in web sources.
Update an enrichment's description or format on a Webset.
Search the web using Exa's AI-native search engine. Supports neural search, auto mode, fast search, and deep search with query expansion. Use **categories** to focus on specific content types like companies, people, news, tweets, or research papers. Filter results by domains, dates, and text content. Optionally retrieve full text, highlights, and summaries inline.
Read the API key’s native team context and Websets concurrency limits. This identifies the team, not a person; product access depends on the key and plan.