Connect once
Use the no-auth Light MCP endpoint for the fastest first run.
Contact
Open navigation menu
Claude + MCP
Turn Claude into a website risk reviewer.
Ask a plain-English question. Get an evidence-backed review of public website behavior, including privacy, cookies, consent, accessibility, policy, and disclosure signals.
No account to startPublic websitesEvidence-linked results
Scan your own website
Paste after connecting CertScore.ai to Claude
Ask me for the public website URL I want to review, and wait for my reply. Then use CertScore.ai to scan only that URL within existing access limits, allowing an eligible recent result to be reused. Request a fresh scan only if I explicitly ask. Retrieve the completed CertScore.ai report and check returned versus total findings and truncation. Recover missing findings with supported bounded retained-report reads for the same scanId, following returned byte-budget guidance and cursors within access limits; use at most five evidence pages and stop on completion, unavailable data or errors. Never start a new scan to fill report gaps. State any findings still unavailable. Group overlapping findings and give three distinct evidence-linked next actions when supported, or fewer if evidence is insufficient. Keep network-request timing, cookie snapshots, new cookie-write timing, completed consent clicks and confirmed consent decisions separate. Include fresh versus reused provenance (or unknown), original scan time, target, region, scope, material coverage limits and the report link. Results are observations, not legal conclusions.
Claude will ask for your URL before scanning. Start with one public page and review its coverage before choosing another website.
Optional: verify the connection with a retained example
This reads an existing example report, preserving its original date and coverage. It does not create a scan or describe your website.
List the CertScore.ai Light tools and read certscore://example-report. Retrieve its retained bundle if needed, preserving the original date and coverage. If unavailable, say so; do not create a replacement scan.
A faster first pass
From “check this site” to a reviewable result.
CertScore.ai gives Claude a structured way to start the work, while keeping evidence, limitations, and report links visible in the response.
Supply your website URL
Copy the website prompt above, paste it into Claude, then provide the public URL you want to review.
Review evidence
Get a bounded result with findings, references, coverage limits, and a report URL.
Light: start without an account
Add a custom remote MCP connector in Claude and paste this endpoint. Light requires no CertScore.ai account, API key or OAuth and provides low-volume public scans with public reports. Claude account, plan and organization requirements are separate.
https://mcp.certscore.ai/mcp/lightLight tools, workflow, and current limitsClaude’s custom connector instructionsHosted OAuth: connect your workspace
Use authenticated access to retrieve reports and previous scans in your authorized CertScore.ai workspace. An eligible account, active workspace membership, and a supported OAuth client are required. Existing access limits apply.
Set up workspace accessAfter your first report
Review another client, brand, or vendor website
Keep the same conversation so both report links stay at hand. Copy this prompt when you are ready to supply a second website; each report retains its own scope and evidence.
Ask me for a different public website URL to review next, and wait for my reply. Use CertScore.ai for only that URL with the same settings where supported, allowing eligible reuse; request fresh observations only if I explicitly ask. Keep this review separate from the first website. Retrieve the completed CertScore.ai report and check returned versus total findings and truncation. Recover missing findings with supported bounded retained-report reads for the same scanId, following returned byte-budget guidance and cursors within access limits; use at most five evidence pages and stop on completion, unavailable data or errors. Never start a new scan to fill report gaps. State any findings still unavailable. Group overlapping findings and give three distinct evidence-linked next actions when supported, or fewer if evidence is insufficient. Keep network-request timing, cookie snapshots, new cookie-write timing, completed consent clicks and confirmed consent decisions separate. Include fresh versus reused provenance (or unknown), original scan time, target, region, scope, material coverage limits and the report link. Results are observations, not legal conclusions.
Prompt library
Useful questions to ask Claude
“Scan my website for cookies and trackers observed before consent.”
“Review this domain before vendor onboarding and summarize the evidence-backed concerns.”
“Using the two completed reports already in this conversation, compare their observed privacy and consent signals. Keep scan dates, settings, and coverage differences visible; do not start new scans.”
“Find the most important accessibility, cookie, policy, and disclosure signals on this site.”
Agencies
Turn client-site reviews into a repeatable workflow your team can run from a Claude conversation.
Privacy teams
Get a fast public-web evidence pass before deeper policy, vendor, or consent review.
Developers
Use a remote MCP server with clear tools, bounded results, and copyable setup instructions.
Before you connect
Frequently asked questions
Do I need a CertScore.ai account?
Light MCP does not require an account, API key, bearer token, or OAuth. It is for low-volume public scans with public reports. Hosted OAuth connects an eligible CertScore.ai workspace for authorized scan history and report access; account and workspace requirements apply.
What can Claude scan?
CertScore.ai scans public HTTP or HTTPS websites for observable accessibility, privacy, cookie, consent, policy, and disclosure risk signals.
What should I ask Claude first?
Paste this prompt after connecting: Ask me for the public website URL I want to review, and wait for my reply. Then use CertScore.ai to scan only that URL within existing access limits, allowing an eligible recent result to be reused. Request a fresh scan only if I explicitly ask. Retrieve the completed CertScore.ai report and check returned versus total findings and truncation. Recover missing findings with supported bounded retained-report reads for the same scanId, following returned byte-budget guidance and cursors within access limits; use at most five evidence pages and stop on completion, unavailable data or errors. Never start a new scan to fill report gaps. State any findings still unavailable. Group overlapping findings and give three distinct evidence-linked next actions when supported, or fewer if evidence is insufficient. Keep network-request timing, cookie snapshots, new cookie-write timing, completed consent clicks and confirmed consent decisions separate. Include fresh versus reused provenance (or unknown), original scan time, target, region, scope, material coverage limits and the report link. Results are observations, not legal conclusions.
How do I review a second website?
After reviewing the first report, copy the next-website prompt and supply a different public URL when Claude asks. Each target is a separate scan request under the existing limits. Light access is not an unlimited bulk-scanning service.
Are results legal or compliance determinations?
No. Results are automated public-web observations with coverage limitations. Review retained evidence and applicable context before relying on a finding.
Make the first call count
Connect CertScore.ai to Claude and scan a real site.
Connect through Light or your eligible workspace, copy the website prompt, and give Claude the public URL you want to review.