Hexowatch
Developer ToolsCreate and manage Hexowatch monitors that watch any web page and report when it changes. Agents react to a competitor price change or a policy edit the moment it is detected.
What This Integration Enables
Hexowatch is for the pages you need to know about and cannot instrument. You do not own a competitor's pricing page, a regulator's guidance page, or a supplier's terms document, so there is no webhook to subscribe to and no API to poll. Hexowatch checks them for you on a schedule and reports when they move. What makes it more than a diffing tool is that it asks you to declare what "changed" means before it starts. A visual monitor compares pixels. A content monitor compares readable text and can be pointed at specific phrases. A keyword monitor fires when a term appears or disappears. A source code monitor compares raw HTML, CSS or JavaScript. A technology monitor fingerprints the stack and, unlike the rest, distinguishes added, updated and removed entries. Availability, domain WHOIS and backlink monitors round out the set.
FlowRunner agents create those monitors from inside a flow, pause and resume them, force an out of schedule check, and read back the full before and after pair for any detection. That last part is what turns a monitor into an input a workflow can reason about instead of an email somebody skims. The agent does the watching and the retrieval continuously. It hands the interpretation to a person at the point where a wrong reading would cost something, which is the whole premise of Orchestration as a Service: agents that know when to stop and ask.
Without FlowRunner
With FlowRunner
Use Case Scenarios
Competitive pricing watch that does not cry wolf
A revenue team wants to know when a named competitor moves list price. The agent creates a content monitor with the price block's phrasing listed explicitly, so only that passage is compared, and a visual monitor on the same URL as a backstop for a redesign that breaks the text selector. When On Change Detected fires, the agent calls Get Scan Result, posts the old and new values into Slack with the change percentage, and waits. If the pricing owner confirms it is a real move, the agent writes the new figure and the capture timestamp into the competitive tracker in Google Sheets and opens a repricing review. If the owner marks it cosmetic, the agent raises the monitor's change level so the same banner rotation stops firing.
Guarding your own pages against silent edits
A source code monitor watches the checkout page's raw markup for the analytics and payment scripts that are supposed to be there. A technology monitor watches the same host's stack. When either reports a change that nobody scheduled, the agent pulls the scan result, opens a ticket in Jira with the exact fragment that differs, and pages the on call channel. The value here is not that a script changed. It is that the change was noticed by something that was not the person who made it.
Domain expiry and supplier stack drift
A WHOIS monitor tracks registration and expiry data for the domains a company depends on, with an alert day set ahead of expiry. A separate technology monitor watches a key supplier's public site. When the supplier swaps a payment provider or a hosting stack, the agent records the added and removed technologies in the vendor record in Airtable and notifies the account owner. Neither of these is an incident. Both are the kind of quiet change that only becomes visible in a renewal conversation months later.
Human-in-Loop Highlight
A visual monitor reports that 6 percent of the page changed. That number is a measurement, not a meaning. Six percent is a price cut and it is also a rotated hero image, and Hexowatch cannot tell you which, because pixels do not carry intent. So the agent does not act on the percentage. It calls Get Scan Result, posts the previous and current screenshots into the pricing channel with the region it believes moved, and asks: "Competitor pricing page changed 6 percent at 04:15. Is this a list price move, or a promotional banner? Confirm to update the tracker and open a repricing review." The human-in-the-loop step here is not ceremony. Repricing on a misread banner is a real revenue decision made on a bad input.
The same restraint applies to spend. Check Monitors Now with an empty selection forces an immediate check of every monitor on the account and consumes credits for each one, and Hexowatch publishes no credit balance endpoint, so the agent cannot verify headroom before committing. An account wide refresh is therefore always an explicit human instruction with the monitor list named, never a default the agent falls back to.
Agent Capabilities
15 actionsMonitor Creation
8- Create Visual Monitor Creates a monitor that screenshots the address on a schedule and reports pixel level differences between consecutive checks. Choose the emulated device to control the viewport and set a change level so small cosmetic shifts do not raise alerts. Use it when the page's meaning is carried by layout rather than text.
- Create Content Monitor Creates a monitor that extracts readable text and reports wording changes. Watch the whole page or list specific phrases so only those passages are compared. This is the right choice for prices, stock wording, terms of service and job listings.
- Create Keyword Monitor Creates a monitor that alerts when supplied keywords appear on or disappear from the address. Match Any Keyword fires on the first hit, Match All Keywords requires every term. Used for back in stock watches, competitor product names and compliance wording.
- Create Source Code Monitor Creates a monitor that compares the raw HTML, CSS or JavaScript served by the address. Watch the complete file or supply specific snippets. Used to detect tracking script changes, removed schema markup and unauthorised code injection.
- Create Technology Monitor Creates a monitor that fingerprints the software the address runs on and alerts when the stack changes. This is the only monitor type that distinguishes added, updated and removed technologies, which makes it the one to use for vendor change detection.
- Create Availability Monitor Creates a monitor that repeatedly requests the address and alerts when the site becomes unreachable or returns an error. It has no content settings, so pair it with a short check frequency for uptime coverage of critical pages and endpoints.
- Create Domain WHOIS Monitor Creates a monitor that tracks domain registration data including owner, status, DNS records and expiry dates. Set an alert day to be warned a fixed number of days before or after expiry. Used to guard your own domains and to watch competitor or acquisition target domains.
- Create Backlink Monitor Creates a monitor that tracks links pointing at the address for a given keyword and alerts when backlinks are gained or lost. Used to police link building campaigns and spot removed partner links.
Monitor Control
4- List Monitors Retrieves every monitor on the account with its monitoring ID, watched address, display name and active state. This is how an agent discovers the monitoring IDs the log, scan result, pause, resume and check actions require.
- Pause Monitors Suspends the selected monitors so checks stop and credits stop being consumed, while configuration and history stay intact. An empty monitor list pauses every monitor on the account, so agents should always name the monitors explicitly.
- Resume Monitors Reactivates paused monitors so checking resumes on the configured schedule. An empty monitor list resumes every monitor on the account.
- Check Monitors Now Forces an immediate out of schedule check of the selected monitors, useful after a deployment or when a flow needs a fresh comparison on demand. The check runs asynchronously, so poll Get Monitor Logs afterwards for the result.
Change History
2- Get Monitor Logs Retrieves a monitor's check history newest first, with the scan result ID, timestamp, whether a change was detected and how much of the page changed. By default only changed checks are returned; turn that off to see clean checks too.
- Get Scan Result Retrieves the full detail of one check as a newData and oldData pair, so a flow can compare exactly what changed. The shape of the pair depends on the monitor type, so match the type reported by Get Monitor Logs.
Notifications
1- List Notification Integrations Retrieves the Slack, Telegram, Discord and email destinations configured on the account, with the integration ID needed when creating a monitor. Destinations are created in the Hexowatch dashboard; this action reads them so a flow can attach existing ones to new monitors.
Triggers
1 triggersEvent Triggers
1- On Change Detected Watches one monitor and fires once for every check that detected a change since the previous cycle, carrying the scan result ID, timestamp, how much of the page changed and the monitor type. Pass the scan result ID to Get Scan Result to read the before and after data. The first cycle records current history only, so switching the trigger on never replays a backlog.
Frequently Asked Questions
What can FlowRunner do with Hexowatch?
FlowRunner agents can run List Monitors, Create Visual Monitor, and Create Content Monitor in Hexowatch, plus 12 more actions.
Does connecting Hexowatch to FlowRunner require OAuth?
No. Hexowatch connects to FlowRunner with an API key, no OAuth flow required.
Can Hexowatch trigger a FlowRunner workflow automatically?
Yes. Hexowatch supports 1 trigger that can start a FlowRunner workflow automatically.
Start building with Hexowatch
$100 in credits. No card required. Connect in minutes.