← Blog
Product6 min read

Monitors: give your AI agent a memory of your site

OpenRush Monitors run your checks on a schedule and save every result, so your agents start with your site's full history instead of one snapshot.

OpenRush Team
Marketing intelligence for AI agents

OpenRush Monitors run your tools on a schedule and save every result. Your reports stay current, and every agent you run starts with your site's full history instead of a single snapshot.

Agents need history, and most marketing tools only show today

A marketing dashboard shows the current state of a site. A person fills in the rest from memory: we rewrote the titles in March, a competitor launched a comparison page, traffic dipped the week after. An agent has none of that. Each session starts cold, so when you ask why clicks fell, it can describe the numbers it sees now and not much else.

We think this is the main gap for agents doing marketing work. Pulling data is not the hard part. The hard part is knowing what you tried, what moved afterward, and what didn't work, and without that an agent's advice is only as good as one day's numbers. Monitors are our start on closing it: they record how your site and your competitors change, so any agent can read the past before it acts.

What a monitor does

A monitor runs the same check on a schedule, on OpenRush's servers, and saves the result. After each check, OpenRush works out what changed since the last one. Your laptop can be off.

Each monitor asks one fixed question, so every check compares fairly with the one before it. To ask something different, create a new monitor and pause the old one, which keeps its history readable.

The three questions monitors answer

What changed this week?

See which searches and pages gained or lost clicks, plus new site problems and lost backlinks. Your agent reads the ranked changes and writes the summary.

Are competitors or AI answers overtaking us?

See how your numbers compare with your rivals' over time, and how often Google AI Overviews cite you compared with them. AI visibility covers Google AI Overviews only, from a sampled corpus, so read trends as directional.

What happened before?

Ask about any week since you started, even if nobody looked at the time. For example:

Show me how /pricing has moved over the last 3 months.

Reading monitor history is free, so your agent can ask as many questions of the past as it needs.

Log what you changed, so agents can connect cause and effect

History of numbers only answers half the question. The other half is what you did. log_change records a change as soon as it goes live, such as rewritten titles, a published page, a technical fix, or earned links. The change then appears by date in get_monitor_history, next to the results.

That lets an agent read a move as what followed an action. It won't claim the action caused it, because a ranking can move for reasons nobody logged. Over a few months, though, you build a record of which changes were followed by gains and which weren't, and the next agent starts from that record rather than from zero. You can ask:

Which title changes we made this quarter were followed by more clicks, and which weren't?

Monitor types

Pick the type that matches the question you want answered:

TypeWhat it tracksHow oftenNeeds
Search ConsoleClicks, impressions, and position over the last 7 complete days, plus top queries and pages over 7 and 28 daysDaily, weekly, monthlySearch Console connected
Website analyticsVisitors, conversions, channels, and landing pages over the last 7 and 28 complete daysDaily, weekly, monthlyGoogle Analytics connected
Site healthA technical audit of your pages, plus new pages foundWeekly, monthlyNothing
RankingsPositions for up to 100 keywords in one country. For a city, up to 10 keywords checked live, with your Google Maps spotWeekly, monthlyNothing
BacklinksLink authority and referring domains gained and lostWeekly, monthlyNothing
CompetitorsDomain numbers for your site and up to 5 competitorsWeekly, monthlyNothing
AI visibilityGoogle AI Overview mentions compared with up to 9 competitorsWeekly, monthlyNothing
AdsSpend, clicks, conversions, and ROAS over the last 7 complete daysDaily, weekly, monthlyA connected ad account

Google data needs a one-time consent, accepted by a signed-in team member. Monitors are shared across your account, and every member can edit, pause, or delete them.

Reports and cost

OpenRush runs the checks and stores the results. It doesn't send reports itself: your agent reads what changed and sends a summary to Slack or email, for example from a scheduled task in your agent's host. Because results are saved, each new agent starts from past results instead of a single snapshot.

A check costs the normal credit price of each tool it runs. There's no extra monitoring fee.

ActionCost
Running a checkThe normal credit price of each tool it runs
Creating, listing, changing, and deleting monitorsFree
Reading monitor historyFree
Logging a changeFree

create_monitor returns the credits per check and per month before you commit, and list_monitors shows how many weeks your balance lasts. A check never charges into a negative balance. If your balance can't cover the next check, the monitor pauses, keeps its history, and resumes once you add credits. Turn on auto-recharge in billing so your monitors keep running.

Set up your first monitor

You can start from the dashboard or from your agent:

  1. Open Monitoring in your dashboard, click Monitor, and pick a site and a type.
  2. Copy the ready-made prompt into your agent.
  3. Your agent suggests any keywords or competitors, states the cost, and creates the monitor once you agree.

Or ask your agent directly:

Use OpenRush to monitor example.com and my three main competitors every week, and send me a summary of what changed in Slack.

The Use Monitors guide covers each type in detail. If your agent isn't connected yet, follow the MCP setup guide.

Also in this update

  • Auto-recharge: turn it on in billing so monitors don't pause when credits run low.
  • Credit costs: prices rose on 15 tools, including backlink gap analysis. Full prices are on the credits page.
  • Privacy policy and terms: we updated both. Please review the privacy policy and terms.

Start before you need the answer

The history builds whether or not anyone reads it, so a monitor started today can answer questions about weeks that haven't happened yet. We think agents will do more of the marketing work, and they will do it better with a record of what happened, what you changed, and what moved afterward. Monitors are where that record starts.

Frequently asked questions

What are OpenRush Monitors?
A monitor runs the same OpenRush check on a schedule, saves each result, and works out what changed since the last check. Your agent reads that history to explain what moved, and you can ask about any week since the monitor started.
How much do OpenRush Monitors cost?
There is no monitoring fee. A check costs the normal credit price of each tool it runs. Creating, listing, changing, and deleting monitors, reading history, and logging changes are free. create_monitor returns the credits per check and per month before you commit.
Does OpenRush send the monitor reports?
No. OpenRush runs the checks on its servers and stores the results, so your laptop can be off. Your agent reads what changed and sends the summary to Slack, email, or wherever you want it, for example from a scheduled task.
Can I change what a monitor checks?
You can rename a monitor or change how often it runs, but not what it asks. A fixed question is what makes checks comparable over time. To ask something different, create a new monitor and pause the old one to keep its history readable.
How do I tell my agent what I changed on my site?
Use log_change. Your agent records a change such as rewritten titles or a published page, and it shows up by date in get_monitor_history next to the numbers. It's free. Describe the change and what it touched, without metrics from connected accounts or personal data.
Share

Give your agent live search data.

One-click setup with Claude or any MCP-compatible agent.