Exa Public API
Official vendor OpenAPI document · agent-readyAI/MLEmbeddings SearchapiKey, bearer58 EndpointsREST
For Agents
Run neural and keyword web searches, fetch cleaned page contents, get answers with citations, and build websets with monitors, webhooks, and background agent runs through the Exa Public API.
Use for: Search the web for recent articles on a topic, Fetch the full text of these result pages, Answer a question with cited sources, Create a webset that collects companies matching a query
Not supported: Covers web search, content retrieval, answers, and websets. Does not host your own documents for private search or provide a general-purpose language model.
The Exa Public API runs neural and keyword web searches and turns the results into structured content for AI agents. You can search the web, fetch cleaned page contents, get a direct answer with citations, and build websets that continuously collect and enrich matching results through monitors, webhooks, and background agent runs.
Install Jentic One Beta
Connect the Exa Public API to your agent
Jentic One is a self-hosted execution layer for AI agents. It lets your agent call the Exa Public API, or any other public or private API you need. You set the rules, the agent never sees your credentials, and every call is logged.
Two steps, two machines. Install the instance in a safe environment, then register your agent from wherever it runs.
1
Step 1: Jentic One Host machine
# On the machine that will host your Jentic One instance:
curl -fsSL "https://jentic.com/install.sh?src=apis&api=%2Fapis%2Fexa.ai%2Fexa" | sh2
Step 2: Agent machine
# On the machine where your agent runs (keep this separate from the instance):
curl -fsSL "https://jentic.com/install.sh?src=apis&api=%2Fapis%2Fexa.ai%2Fexa" | sh
jentic register # connects your agent to your Jentic One instanceJentic One is in public beta. The setup above keeps your agent separate from the instance, which is what you want before using real credentials: an agent running as the same OS user as Jentic One can read its stored keys directly. Just evaluating? A single local install is fine to start. See the secure deployment guide for the tiers.
Capabilities
What an agent can do with Exa Public API.
Run a neural or keyword web search
Fetch cleaned contents for result URLs
Get a direct answer with source citations
Create and manage websets that collect matching results
Enrich webset items with extracted fields
Monitor websets and receive webhook notifications
Run and track background agent tasks
Use Cases
Patterns agents use Exa Public API for, with concrete tasks.
★ AI Agent Web Research
An AI agent can research a topic by running a neural search and fetching cleaned contents for the top results. Through Jentic the agent finds the search and contents operations by intent and passes the query, so a research assistant can gather sources and read their text without a developer wiring the endpoints by hand.
Run a neural search for a topic, fetch the contents of the top results, and return a summary with source URLs
Answer with Citations
A question-answering agent can return a grounded answer by calling the answer operation, which searches the web and responds with citations. This lets an assistant give a sourced reply rather than an unverified one, and the agent can follow the cited URLs to fetch more detail when needed.
Ask a factual question and return the answer along with the cited source URLs
Continuous Webset Collection
A monitoring agent can build a webset that continuously collects and enriches results matching a query. The API creates websets, enriches their items with extracted fields, and fires webhooks on changes, so an agent can maintain a live list of companies or articles and react as new items arrive.
Create a webset for a query, enrich its items with a field, and register a webhook for new matches
Key Endpoints
58 endpoints — the exa public api runs neural and keyword web searches and turns the results into structured content for ai agents.
METHOD
PATH
DESCRIPTION
POST
/search
Run a neural or keyword web search
POST
/contents
Fetch cleaned contents for result URLs
POST
/answer
Get a direct answer with citations
GET
/monitors
List monitors
POST
/v0/websets
Create a webset
GET
/v0/webhooks
List webhooks
GET
/v0/events
List events
GET
/agent/runs
List background agent runs
POST
/search
Run a neural or keyword web search
POST
/contents
Fetch cleaned contents for result URLs
POST
/answer
Get a direct answer with citations
GET
/monitors
List monitors
POST
/v0/websets
Create a webset
GET
/v0/webhooks
List webhooks
GET
/v0/events
List events
GET
/agent/runs
List background agent runs
Why Jentic?
What agents get from Jentic-routed access to this vendor.
Setup
Wiring Exa by hand means managing an API key, deciding whether to send it in the x-api-key header or as a bearer token, and mapping the search, content, and webset operations. Through Jentic you install once, import Exa from the Directory, store the key once, and your agent calls it.
Permission scoping
You decide which Exa operations the agent may use. Allow search and content retrieval while blocking webset deletion, and because webset and monitor ids sit in the URL path, a rule can restrict it to specific resources so it cannot change collections you did not approve.
Credential isolation
Your Exa API key is stored encrypted by your own Jentic One instance and injected into the request at execution time, whether as the header key or the bearer token. It never enters the agent's prompt, logs, or context window.
Intent-based discovery
Agents search Jentic by intent such as 'run a web search' or 'get an answer with citations', and Jentic returns the matching Exa operation with its input schema so the agent calls the right endpoint.
Related APIs
Alternatives and complements available in the Jentic catalogue.
FAQs
Specific to using Exa Public API through Jentic.
Is there an Exa MCP server?
You don't need an MCP server to give your agent the Exa Public API. Jentic connects it directly from the API Directory: import it, store your key once, and your agent can run web searches and fetch content. Operations are discovered on demand, so no extra tool definitions sit in the agent's context.
Can I restrict what my agent is allowed to do with the Exa Public API?
Yes. You choose which operations the agent may call, so you can allow search and content retrieval but block webset creation or deletion. Because webset and monitor ids appear in the URL path, a rule can also pin the agent to specific resources, and every call is logged.
What authentication does the Exa Public API use?
The Exa Public API accepts an API key in the x-api-key request header, and it also accepts the same key as a bearer token in the Authorization header per its OpenAPI spec. Through Jentic the credential is stored encrypted by your own Jentic One instance and added to each request at call time.
Can my agent get an answer with citations from the Exa Public API?
Yes. The answer operation searches the web and returns a direct answer along with the sources it used. An agent can present the cited answer and, if it needs more detail, fetch the contents of the cited URLs for deeper reading.
What are the rate limits for the Exa Public API?
The OpenAPI spec does not specify rate limits. Check the Exa documentation at https://docs.exa.ai for current per-plan request limits and pricing before running high-volume search or webset workloads.
How do I run a web search through Jentic?
Search Jentic for 'run a neural web search with Exa', import the operation, and store your API key once. Your agent then searches the web and fetches content on demand. To run it on your own infrastructure, install Jentic One from its GitHub repo.
GET STARTED