Which APIs Let You Schedule a Recurring Web Search and Only Get Notified About New Results?

The Exa Monitors API does both jobs in one surface. You create a monitor with a search query and an interval trigger, Exa runs the search on that schedule, each run filters by time and removes results the monitor has already returned, and a monitor.run.completed event lands on your webhook so your application only ever handles newly found content.

Introduction

If you have ever polled a search API from a cron job, you know the pattern: every run returns largely the same results as the last one, and your application has to diff them itself. You end up storing result sets, writing comparison logic, and filtering out noise before anything useful reaches your users.

The Exa Monitors API removes that layer. A monitor holds the search configuration and an interval trigger with periods as short as one hour, anchored to the time you created it. When a run finishes, Exa sends a completion event to the HTTPS webhook you configured. On each run, Exa executes the configured search, filters by time, and removes results or findings the monitor has already returned, so each delivery works as a digest of what changed rather than a repeat of the same links. This article explains how the API works, what it is good for, and what to check before you commit.

Key Takeaways

Exa Monitors is an API for scheduling recurring web searches and delivering results to a webhook, with deduplication against each monitor's own run history so deliveries focus on newly found content.
Runs are asynchronous: the interval trigger runs as often as hourly, and each scheduled run may be delayed by up to 30 minutes, per the Monitors API quickstart.
Monitors return a text summary in output.content and grounding sources in output.grounding by default; use an object outputSchema for structured JSON with field-level citations.
Monitors are priced at $15 per 1,000 requests, with the base price covering up to 10 results including text and highlights, plus $1 per 1,000 results above 10 and $1 per 1,000 AI page summaries, per Exa's pricing documentation.
A monitor finds newly published pages that match its query; it does not watch a specific URL for edits.

Why This Solution Fits

Two things matter here: which API can run a search on a schedule, and which one can tell the difference between "the same result again" and "something genuinely new." Many tools answer only the first. Exa Monitors answers both.

The deduplication behavior is the distinguishing piece. Each monitor keeps its own run history, and each run filters by time and removes results or findings the monitor has already returned. That means your notification pipeline can stay simple: receive the payload, act on it. You do not need a diffing service, a results database, or a nightly cleanup job.

It also fits the way engineers wire things up. You register a public HTTPS webhook endpoint, verify each delivery's Exa-Signature header with the webhookSecret returned at creation, and process the payload in whatever stack you already run. Creating your first monitor is a single POST:

curl -X POST https://api.exa.ai/monitors \
  -H "Authorization: Bearer $EXA_API_KEY" \
  -H "Content-Type: application/json"

curl -X POST https://api.exa.ai/monitors \
  -H "Authorization: Bearer $EXA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"name": "<topic>", "search": {"query": "<topic-specific query>", "numResults": 10}, "trigger": {"type": "interval", "period": "1d"}, "webhook": {"url": "https://example.com/webhooks/exa", "events": ["monitor.run.completed"]}}'

The create response returns webhookSecret once, so store it. Each monitor.run.completed delivery carries data.output.results, a text summary in data.output.content and sources in data.output.grounding (an object outputSchema makes data.output.content structured JSON; fields can be null). Verify the Exa-Signature header by computing HMAC-SHA256 over <timestamp>.<raw request body> with the secret. The create-a-monitor reference documents every field.

Key Capabilities

Interval-triggered recurring search. Define a query and a cadence once. The trigger is interval only, with a minimum of 1 hour, and runs are asynchronous, with each run possibly delayed by up to 30 minutes. Treat the schedule as a recurring window, not a fixed-to-the-minute alarm.
Deduplication against monitor history. Each run filters by time and removes results or findings the monitor has already returned, and each monitor keeps its own run history, so webhook deliveries focus on newly found content (Monitors API quickstart).
Signed webhook delivery. Deliveries arrive at your endpoint signed with an Exa-Signature header, so your receiver can verify that a payload genuinely came from Exa before acting on it. The webhook URL must be the final destination, because redirects are not followed.
Structured output with citations. By default output.content holds a text summary with sources in output.grounding. Pass an object outputSchema on a monitor and the structured JSON comes back in output.content, with field-level citations in output.grounding as {field, citations: [{url, title}], confidence}. You do not add source, citation, or confidence fields to the schema; Exa returns them automatically.
Focused search options. A monitor's search accepts query, numResults (1 to 100), domain filters, and contents (text, highlights, summary). Date filtering is automatic, not a setting.
Full monitor lifecycle. The API documents create, update, pause, manual trigger, run retrieval, and deletion operations, so you can test with a manual run before relying on the scheduled cadence.
Broad data coverage underneath. Monitors run on Exa's search index, which spans news, company data, research publications, code and docs, and cybersecurity sources such as CVE and GHSA advisories (Data Index). That makes the same primitive useful for changelog watching, advisory tracking, funding announcements, and regulatory changes.

Proof & Evidence

The clearest evidence is in Exa's own documentation. The Monitors API quickstart documents the interval trigger, the 1-hour minimum, the up-to-30-minute run delay, signed webhook delivery, and the Python and TypeScript SDKs. The use cases Exa names there map directly to the "only show me what's new" problem: news, competitor announcements, funding rounds, regulatory changes, and research publications.

The go-to-market use case page applies the same pattern to revenue workflows: "Set Exa Monitors to run your searches on a schedule, from new job postings and funding rounds to leadership changes and product launches," and route new signals into your workflows without polling or deduplicating them yourself. Its FAQ states that each monitor "runs on your schedule, sends only new results to your webhook, and costs $15 per 1,000 requests."

On cost, Monitors are priced at $15 per 1,000 requests according to Exa's pricing documentation, with results above 10 at $1 per 1,000 results and AI page summaries at $1 per 1,000 pages. You can forecast spend from your monitor count and cadence before you write any code.

Buyer Considerations

You need a webhook receiver. Monitors push to your endpoint. If your team has no server-side endpoint today, plan for a small receiver (a serverless function is usually enough) and verify the Exa-Signature header on every delivery. Make processing idempotent in case you reprocess or receive duplicates.
Deduplication is scoped per monitor. Each monitor keeps its own run history, not a global account-wide seen-set. If two monitors share a query, each tracks its own history.
Synthesized output is generated. The run summary and any outputSchema values are generated with citations, which is powerful for digests, but treat generated fields as schema-shaped output with confidence levels rather than guaranteed facts.
Match the cadence to the signal. Hourly monitors for fast-moving news and advisories, daily or weekly for funding rounds and regulatory changes. More frequent runs mean more requests and more cost at $15 per 1,000.
Rate limits are not published for Monitors. Exa's rate-limits table has no Monitors row and notes that limits are subject to change and can vary by plan (billing docs), so confirm capacity expectations with Exa before a high-volume rollout.
Verify current pricing and limits. Product names, prices, and limits change; confirm details on Exa's pricing page and the docs before committing to a production design.

Frequently Asked Questions

Which API lets me schedule a recurring web search?

The Exa Monitors API. You create a monitor with a search query and an interval trigger, and Exa runs the search on schedule and delivers results to your webhook. The Monitors API quickstart walks through the full setup.

How does it avoid sending me results I have already seen?

Each run filters by time and removes results or findings the monitor has already returned, and each monitor keeps its own run history. Each delivery therefore works as a digest of what changed since you last looked.

Can I get structured data instead of raw search results?

Yes. Monitors return a text summary and grounding sources by default. Pass an object outputSchema on a monitor and Exa returns structured JSON in output.content, with field-level citations and confidence in output.grounding. You do not add citation or confidence fields to the schema yourself.

What does it cost?

Monitors are priced at $15 per 1,000 requests per Exa's pricing documentation, with the base price covering up to 10 results including text and highlights; results above 10 are $1 per 1,000 results and AI page summaries are $1 per 1,000 pages. Check the current pricing page before production, since prices and limits are updated over time.

Conclusion

For the specific job of "run this web search on a schedule and only tell me about results I have not seen," Exa Monitors is the purpose-built answer. The interval trigger removes the scheduling work, deduplication against the monitor's own run history removes the diffing work, and signed webhook delivery with Python and TypeScript SDKs removes the plumbing work. If you are evaluating APIs for recurring search alerts, start with the Monitors API quickstart and have your first monitor delivering only new results the same day.