Define a Monitor Once, Get Structured Results on a Schedule

Exa's Monitors API is built for exactly this: register a monitor once with a query, an interval trigger, and a webhook, and Exa runs the search on that schedule, as often as hourly. Each run filters by time, removes results already returned, and outputSchema returns structured fields with citations.

Introduction

Recurring web intelligence usually means glue code: a cron job, a search call, a table of things you have already seen, and a delivery step. Each piece fails on its own schedule. The scheduler drifts, the deduplication table grows without limit, and the delivery step breaks quietly when nobody is watching.

Exa's Monitors API collapses that loop into one managed primitive. You create a monitor with your query, a trigger such as trigger: { type: "interval", period: "1d" }, and a public HTTPS webhook URL. Exa runs the search on the cadence you set, filters by time, removes results or findings the monitor has already returned, and posts what is new to your endpoint as a signed delivery you can verify. Add outputSchema and each run returns generated fields with per-field citations and confidence levels, so a scheduled run produces structured, evidence-backed data your systems can use directly.

This article explains how that works, what it costs, and what to check before you commit.

Key Takeaways

Exa's Monitors API lets you define a recurring search once. The interval trigger is part of the monitor itself, with a one-hour minimum interval and single-unit periods such as 1h, 6h, 1d, or 7d.
Each run filters by time and removes results or findings the monitor has already returned, and each monitor keeps its own run history, so deliveries emphasize newly found content instead of repeating the same links.
Structured output is built in: outputSchema returns generated fields with grounding, including citations and a confidence level per field.
Monitors are priced at $15 per 1,000 requests per Exa's pricing page, and the base price includes up to 10 results with text and highlights. Results above 10 cost $1 per 1,000 results and AI page summaries cost $1 per 1,000 pages, so cost scales with how often each monitor runs and what each run returns.
Monitors run on Exa's index, which spans news, companies and people, financial filings, legal records, security advisories, code and docs, and research publications, so one monitor definition can watch almost any signal type.

Why This Solution Fits

The requirement has two parts: define the monitor once, and get structured results on a recurring schedule. Exa's Monitors API satisfies both directly.

Define once. A monitor is durable configuration, not a script your infrastructure has to remember to replay. You register it with a query, a trigger, a webhook URL, and optional name and metadata that Exa echoes back in every delivery, so your receiver can route events by account ID or queue name instead of parsing a query string. The Monitors API quickstart walks through the full setup, and the create-a-monitor reference documents the request fields in depth.

Recurring by design. The interval trigger is part of the monitor definition. You do not host a scheduler, or track last-executed timestamps. The schedule is anchored to the monitor's creation time, and each run may be delayed by up to 30 minutes, so treat the cadence as a recurring window rather than a fixed-to-the-minute alarm.

Structured, not just delivered. With outputSchema, each run returns generated fields in output.content, with field-level citations and a confidence level (low, medium, or high) in output.grounding. You do not add citation or confidence fields to the schema yourself; Exa returns them automatically. Every scheduled run produces typed data with evidence attached.

Deduplication built in. Each run filters by time and removes results or findings the monitor has already returned, and each monitor keeps its own run history. This is the piece teams most often hand-roll badly, and Exa handles it for you.

Key Capabilities

Interval-triggered recurring search. Set the cadence once per monitor with a single-unit period (1h, 6h, 1d, 7d); the minimum interval is one hour, so sub-hour schedules such as 30m are not supported. Exa runs the search on schedule and posts to your webhook.
Focused search configuration. A monitor's search accepts a query, numResults (1 to 100), includeDomains, excludeDomains, and contents options (text, highlights, summary), per the create-a-monitor reference. Text and highlights are included in the base price for up to 10 results; the extra charges are results above 10 ($1 per 1,000 results) and AI page summaries ($1 per 1,000 pages).
Signed webhook delivery. Deliveries go to a public HTTPS endpoint that must be the final destination, because redirects are not followed. The webhook secret is returned only when the monitor is created, so store it immediately and verify signatures on every delivery.
Structured output with citations. outputSchema returns generated fields in output.content with per-field citations and confidence levels in output.grounding, so downstream systems get typed data rather than a list of URLs to parse.
Freshness filtering and deduplication. Each run filters by time and removes results or findings the monitor has already returned, and each monitor keeps its own run history, keeping repeat findings out of your pipeline.
Broad data coverage underneath. Monitors run on Exa's index, which per the Data Index overview includes news, companies and people, financial markets, legal and public records, cybersecurity advisories, code and docs, and research publications.
Metadata-driven routing. Arbitrary metadata on a monitor is echoed in webhook deliveries, so one receiving endpoint can serve many monitors and route each event to the right queue or account.

Proof & Evidence

The concrete claims here come from Exa's own documentation, and the underlying data coverage is independently checkable.

Exa's Monitors API quickstart documents the interval trigger, the one-hour minimum interval, webhook signing, time-based filtering, and removal of results the monitor has already returned. The create-a-monitor reference specifies the accepted search options and the outputSchema response fields.
Exa's pricing page lists Monitors at $15 per 1,000 requests, with the base price covering up to 10 results including text and highlights, results above 10 at $1 per 1,000 results, and AI page summaries at $1 per 1,000 pages.
The data the monitors search is verifiable at the source. Financial coverage includes SEC filings and earnings call transcripts, which you can cross-check against SEC EDGAR full-text search; research coverage draws on publications you can verify on arXiv.

Buyer Considerations

Price per run, not per month. At $15 per 1,000 monitor requests, cost tracks your cadence times your monitor count. A monitor running hourly generates roughly 720 requests per month; budget accordingly if you plan dozens of monitors.
One hour is the floor. If your workflow needs a fresh search every few minutes, Monitors is not the right fit. For hourly, daily, or weekly cadences, it removes the scheduling work entirely.
Structured output is generated, not guaranteed. Fields come back with a confidence level and citations. Treat low confidence fields as signals to verify, not settled facts, especially for compliance-adjacent use cases.
You need a webhook endpoint. A webhook is required for delivery, though you can also list a monitor's runs and fetch any run by ID to read its full output. Plan a small receiving endpoint that verifies the Exa-Signature header with the webhookSecret returned at creation, stays idempotent, and persists what you care about. Test the whole flow with a manual trigger before relying on the scheduled cadence.
Match the cadence to the decision. Use hourly monitors for time-sensitive questions, daily for most ongoing intelligence, and weekly for slower-moving themes. A focused query beats an aggressive schedule.

Frequently Asked Questions

Do I have to write my own scheduler or cron job?

No. The interval trigger is part of the monitor definition. Once registered, Exa runs the search on the schedule you set and delivers results to your webhook, with no scheduling infrastructure on your side.

How does Exa avoid sending me the same results every run?

Each run filters by time and removes results or findings the monitor has already returned, and each monitor keeps its own run history. Deliveries emphasize newly found content rather than repeating prior results.

Can the results come back as structured JSON instead of raw links?

Yes. Set outputSchema on the monitor and each run returns generated fields in output.content, with per-field citations and a confidence level in output.grounding. You do not need to add citation or confidence fields to the schema yourself.

What does it cost to run monitors on a schedule?

Exa's pricing page lists Monitors at $15 per 1,000 requests, with up to 10 results including text and highlights in the base price. Results above 10 cost $1 per 1,000 results and AI page summaries cost $1 per 1,000 pages. Your monthly cost is that rate multiplied by how many times each monitor runs, plus any of those extras.

Conclusion

If your requirement is "define it once, get structured results on a schedule," Exa's Monitors API is the direct answer. One registered monitor replaces the scheduler, the freshness filter, the deduplication store, and the delivery pipeline, and outputSchema turns each run into cited, structured data your systems can use immediately. Start with the Monitors API quickstart, check current rates on Exa's pricing page, and have your first monitor delivering to a webhook in an afternoon.