Perplexity Monitor API: Power-User Guide to Search Monitors, Page Diffs and Webhooks

Last updated: November 17, 2026

The Perplexity Monitor API, launched November 16, 2026, checks the web on a schedule and saves relevant findings for your app. A Search monitor tracks new results for one query; a Page monitor tracks text changes on up to 10 URLs. Both cost $0.001 per billable unit, and webhooks are optional.

Contents

What the Monitor API does

Perplexity announced the Monitor API on November 16, 2026, describing it as a way to track new developments on a topic, or changes to specific pages, on a schedule and send relevant updates to your application. It was available at launch.

According to the Monitor API overview, a monitor stores state between runs so it can tell what is new. Each run saves kept findings in changes and page errors in target_errors. Monitors can also be set up in the API Console.

Endpoints live under https://api.perplexity.ai/v1/monitors and cover create, update, pause, resume, delete, run now, run history and webhook tests. The Python and TypeScript SDKs expose them as client.monitors.

Search vs Page monitors

Every monitor has either one Search target or 1 to 10 Page targets, never a mix.

Feature Search monitor Page monitor
What it watches New web results for one query Text changes on 1 to 10 known URLs
Target fields query (up to 1,000 characters), max_results (1 to 20, default 20), filters url: public HTTP or HTTPS, up to 2,048 bytes
Filters Freshness (published, updated) and up to 20 included or excluded domains None; choose the URLs
First run Looks back one schedule period, up to seven days Saves a baseline and reports no change
Result type search_result with URL, snippet and confidence page_change with a unified diff and confidence
Price $0.001 per succeeded run $0.001 per counted fetched URL

The Search monitor guide notes that a page edited after it was reported is not reported again. To follow later edits at a known address, the Page monitor guide is the right tool: it compares fetched text with the saved version and returns a diff capped at 24 KiB.

How to create your first monitor

  1. Get an API key. Export a Perplexity API key as PERPLEXITY_API_KEY and make sure the organization has credits.
  2. Choose the type. Use one search target for a topic, or up to 10 page targets for URLs such as release notes.
  3. Write a goal. Add an optional goal, for example “New Series A or B announcements” or “New features and breaking changes; ignore navigation and footer edits.”
  4. Set a schedule. Use {"type": "interval", "interval": "1d"}. The docs list 1h, 6h, 1d and 1w as example intervals.
  5. Send the request. POST the name, targets, goal and schedule to /v1/monitors. Add an Idempotency-Key header if your code retries. The monitor starts as active, and its first run starts right away.
  6. Read the results. Call GET /v1/monitors/{id}/runs?limit=1. Wait while a run is queued or running, then read changes. Do not trigger another manual run for the first check.
  7. Add a webhook if needed. Attach webhook.url, store the one-time secret from the response, and call /webhook/test to confirm delivery.
  8. Pause when you finish testing. A paused monitor keeps its state and stops scheduled runs.

Goals, judging and deduplication

Perplexity runs automatic judging before a finding appears in changes. Search results must pass checks for novelty, source legitimacy and the run’s time window. A relevance check against your goal applies only when you set one.

Page monitors use one meaningfulness check, judged against your goal if you set one. To tune strictness, set judge.threshold on each target, above 0 and up to 1. The default is 0.5; higher values keep fewer changes.

Search monitors deduplicate normalized URLs within a run and against URLs already seen, and changing the goal does not replay old results. Seen-URL history, page baselines and run history are kept for 90 days.

Webhooks and signatures

By default a webhook receives monitor.changes.detected (a run succeeded with at least one change) and monitor.paused (the service paused the monitor). You can also subscribe to monitor.run.completed, which fires for every succeeded or failed run.

  • Signing: deliveries follow Standard Webhooks v1 with webhook-id, webhook-timestamp and webhook-signature headers. Verify the HMAC-SHA256 signature on the raw body and reject timestamps more than five minutes off.
  • Responses: return any 2xx within ten seconds and process the event afterward. Redirects are not followed.
  • Retries: timeouts, network errors, 408, 429 and 5xx responses are retried up to seven attempts over about 17.6 hours. Other 4xx responses and TLS failures end the delivery.
  • Duplicates: delivery is at least once and may be out of order, so deduplicate by webhook-id.

Failing webhooks or sources, insufficient credits or organization policy can pause a monitor automatically. After fixing the cause, call resume.

Pricing and limits

The pricing page sets one rate: $0.001 per succeeded Search run, or per counted fetched URL in a succeeded Page run. Failed and skipped runs are free, and a Search run is billed once even when it returns nothing.

For example, a daily Search monitor runs about 30 times a month, roughly $0.03. Ten pages checked hourly for 30 days is up to 7,200 fetched URLs, about $7.20.

The rate limits page allows 1 request per second on Tier 0, rising to 33 on Tiers 4 and 5. Every Monitor call, including polling, uses that budget. Active monitors are capped at 100 on Tiers 0 and 1, then 200, 300, 400 and 500 on Tiers 2 to 5. Paused monitors do not count, and hitting the cap returns HTTP 409 quota_exceeded.

Monitor API vs the Search API

Before this launch, developers could poll the Perplexity Search API themselves. That costs $5 per 1,000 standard requests, or $1 with Fast Search, and leaves deduplication, change detection and alerting to your code. The Monitor API moves scheduling, state, judging and webhooks to Perplexity at $1 per 1,000 runs.

For a one-off yes or no check on a page, the Decisions API can now fetch a URL; see our Perplexity Decider v1.3 guide. Teams on AWS can compare the AgentCore Web Search API, agent builders can read our Perplexity Agent API guide, and app users can see Perplexity Teammates.

Limitations to know

  • API organizations that require zero data retention cannot create, resume or manually run monitors.
  • Page targets must be public, with no embedded credentials, so logged-in pages are out of scope.
  • State can expire after a pause longer than 90 days, so a Page monitor may start a fresh baseline.

Frequently asked questions

When did Perplexity launch the Monitor API?

Perplexity announced the Monitor API on November 16, 2026, and said it was available immediately at $1 per 1,000 runs or URLs fetched.

How much does the Perplexity Monitor API cost?

A succeeded Search run costs $0.001, or $1 per 1,000 runs, however many searches it performs. Page monitors cost $0.001 per counted fetched URL in a succeeded run. Failed and skipped runs are free.

What is the difference between a Search monitor and a Page monitor?

A Search monitor runs one query on a schedule and saves new, relevant web results. A Page monitor watches 1 to 10 known URLs and reports meaningful text changes as diffs.

Do I need a webhook to use the Monitor API?

No. Webhooks are optional. Without one, runs and their changes stay readable through the API for 90 days.

How many monitors can I run at once?

Active monitors are limited by usage tier: 100 for Tier 0 and Tier 1, then 200, 300, 400 and 500 for Tiers 2 to 5. Paused monitors do not count toward the limit.

Sources

Featured image: Perplexity AI logo, public domain (PD-textlogo), via Wikimedia Commons. Perplexity is a trademark of Perplexity AI, Inc. FreeAnswers is not affiliated with Perplexity.

Related answers

Leave a Comment