AI News
AI News AgentToolingPerplexity3 min read

Perplexity launches Monitor API to watch the web

Perplexity launches `Monitor API`, a tool that monitors topics and web pages to detect relevant updates and changes. It sends alerts through webhooks and lets you connect those updates to agents and automated workflows.

Perplexity has launched Monitor API, a tool that watches the web and alerts applications and AI agents when it finds new information or detects relevant changes on a page.

The idea addresses a practical problem: a company may need to know when a competitor changes its prices, when a startup announces funding, or when an API introduces a change that breaks existing integrations. Until now, many organizations handled this with scheduled searches and custom code to remove duplicates and separate important updates from the noise.

Monitor API automates much of that work and lets you create each monitor using plain language, without having to define complex rules.

Two ways to monitor information

The tool offers two main modes:

  • Monitor the web: searches web and news results for updates about a topic, company, or question.
  • Monitor pages: tracks up to 10 specific URLs, such as a pricing page, changelog, or privacy policy.

In the second case, the first review saves a baseline version. Later reviews compare the page with that version and return a diff, meaning a text summary of what has changed.

Each monitor can include a goal written in natural language. For example: detect new Series A or Series B funding rounds at AI infrastructure companies. You can modify that goal later without creating the monitor from scratch.

Fewer irrelevant alerts

Perplexity uses a system called Decider to review each result before sending a notification. In web mode, it checks four aspects:

  • That the information is relevant to the goal.
  • That the source is legitimate.
  • That it provides new information.
  • That it is recent.

When monitoring pages, the system tries to determine whether a change matters to the defined goal. It only sends the update when the configured thresholds are met and adds a confidence score.

This does not eliminate the need for human oversight. But it can prevent an agent from receiving hundreds of repeated alerts or changes that do not affect its task.

Alerts that trigger actions

Monitors can run every 15 minutes, every hour, every day, or every week. They also accept custom schedules through a cron expression, a common format for scheduling automated tasks.

When it finds an update, the API sends it through a signed webhook. A webhook is an automatic notification that an application receives as soon as something happens. If delivery fails temporarily, Monitor API retries the request.

That lets you connect an alert to a complete workflow. For example:

  • A funding announcement can trigger an investigation into the company and update the CRM.
  • A competitor pricing change can generate a recommendation for the sales team.
  • A regulatory change can be summarized and sent to the compliance team.
  • A breaking API change can open a task for the development team.

You can pause, resume, or run monitors on demand through the API or Perplexity console. They also notify you if they are paused automatically.

Pricing and availability

Web monitors cost $1 per 1,000 successful runs. For page monitors, the price is $1 per 1,000 URLs checked, because a single run can review multiple pages.

Monitor API is available now through the API and Perplexity console, using the same key as its Search, Context, and Agent services.

Perplexity says it tested the tool for eight hours across 60 topics based on production queries, including business news, technology launches, and regulation. According to that evaluation, Monitor returned the most recent and relevant results among the providers compared, and found information that other services did not return.

For you, the important change is not getting another source of alerts. It is being able to turn a new development on the internet into an automated action. The next thing worth watching will be the real-world accuracy of those alerts at scale, especially in areas where a false alarm costs time or money.