Kept AI

Agent

Put your AI search data to work

Ask questions, investigate performance, and move work forward with Kept Agent and connected MCP workflows.

Compatible research and workflow contexts

AI searchCrawler accessProject insightsMCP assistantsTeam reportingContent research

Features

From a quick question to a full workflow

Explore your project in Kept Agent. Bring broader visibility research, reports, and views into your connected assistant through MCP.

Find clear answers in your project data

Ask Kept Agent about crawl activity, crawler access, and recorded conversions, with answers grounded in your project.

Kept Agent

Explore the signals behind a change

Compare overview metrics, prompts, sampled answers, and sources through MCP to find the evidence worth a closer look.

MCP workflow

Bring findings into charts and reports

Read Kept MCP data in your connected assistant, then use that tool to turn findings into a chart, table, or report for your team.

MCP workflow · chart sample

Create focused views for your team

Read your project through MCP, then build a focused view in your connected tool around the questions your team wants to explore.

MCP workflow · view sample

Keep your workspace moving forward

Inspect robots.txt, test a crawler path, and review settings changes with Kept Agent. Owners confirm writes by default and can enable auto-approval.

Kept Agent · owner controls

Bring evidence into your other tools

Read project insights through Kept MCP, then prepare your next brief or report using the connections available in your assistant.

MCP workflow

MCP workflows

Start with a question that matters

Four practical workflows for your connected assistant, grounded in read-only Kept MCP data.

Connect your data to the way your team works

Ask Kept Agent about crawl activity, access, and conversions. For broader research, connect Kept MCP to a compatible assistant and use that tool to prepare charts and reports.

Better decisions start with evidence. Ask a clear question, inspect the source, and make the next move count.

Kept AI
Kept AI editorial principle

Illustrative team use cases

A clearer next step for every team

Sample workflows for research, reporting, and project upkeep.

Inspect a blocked guide with Kept Agent. Test its path for a named crawler and review the matching robots rule before deciding what to change.

Technical SEO

Illustrative team use cases

Turn read-only MCP results into a client report draft in a connected tool. Review its sources and findings with your team before sharing.

Agency lead

Illustrative team use cases

Use MCP sources and citation gaps in a connected assistant to prepare a content review brief. Review the original answers to prioritize the next rewrite.

Content lead

Illustrative team use cases

Compare prompts and sampled answers through MCP. Organize the findings in your connected tool around the questions the team wants to investigate next.

Research operations

Illustrative team use cases

Ask Kept Agent to summarize recorded conversions and crawl activity. Review the project data to shape the next question for your team.

Performance marketing

Illustrative team use cases

Review a gateway robots override for a proxied host or a full conversion goal list with Kept Agent. Confirm the change before it is applied under the default settings.

Project owner

Illustrative team use cases

Your next question starts here

Open your project. Find the evidence. Decide what to do next.

Frequently asked questions

Answers about Kept Agent and your connected AI workflows.

What can Kept Agent help with?

Explore crawl activity, crawler access, and recorded conversions. Inspect robots.txt or test a path for a named crawler. Project owners can also update gateway robots overrides for proxied hosts and replace conversion goal lists.

How do the MCP workflows work?

The four cards show research workflows for a compatible assistant connected to Kept MCP. The assistant reads project data through MCP, then helps you compare findings or prepare a brief in that tool.

How is this different from a general chatbot?

Kept Agent works with crawl, access, and conversion tools in your project. A connected MCP assistant can read permitted project insights, including prompts, sampled answers, and sources, to ground your research.

Can it explain a change in visibility?

Use a connected MCP assistant to investigate overview metrics, prompts, sampled answers, sources, citation gaps, and brand insights. The evidence you can explore depends on available project data; patterns can guide an investigation without proving a cause.

Can I connect other tools?

Connect Kept MCP to a compatible assistant to read your project data. The connections available in that assistant determine where you can use the findings to prepare documents, charts, or reports.

Will it ask before changing settings?

Supported writes require owner confirmation by default, after Kept Agent shows the full change. Owners can enable auto-approval to apply changes without that step. Gateway robots overrides apply only to proxied hosts; conversion goal updates replace the full list.

Can I customize the prompts?

Yes. Ask Kept Agent about a specific host, crawler, or path. For research through MCP, adapt a workflow prompt in your connected assistant to the question you want to explore.

Where should I start?

Sign in to Kept AI, open your project, and ask Kept Agent a crawl or access question. To explore prompts, answers, and sources in a connected assistant, follow the MCP setup guide linked above.

Which plan gives me access?

Check access and project permissions in your Kept AI account, and review current options on the pricing page. To use a connected assistant, follow the MCP setup guide and configure a token with access to your project.

View current plans