Can NoClick check a page's Core Web Vitals automatically?

Yes. The Pagespeed node returns Core Web Vitals for a specific URL from lab tests and Origin Core Web Vitals from real Chrome user data, along with the raw lab metrics behind them. Wire it to a schedule or a deploy trigger and a workflow can track those numbers for any page without anyone opening a browser.

Does it test mobile and desktop separately?

It does. There are separate Analyze (Mobile) and Analyze (Desktop) operations, and Run Analysis (Full) returns the complete report. One workflow can score the same URL on both form factors and compare the results side by side.

How do I connect the Pagespeed node?

The node authenticates with a Google API key for the PageSpeed Insights API, which you add once in the node's credentials. After that any workflow can call it against public URLs, so there is no per-run setup.

Can it tell me what to fix, not just the score?

Yes. Get Opportunities returns the specific improvements PageSpeed suggests for a page, and Get Audit Details gives the underlying data behind each audit. Feed those into an AI step or a ticketing node to turn a low score into an actionable list.

Can a workflow react to a failing score?

The node itself is request and response, so it does not fire on its own, but you can schedule it or trigger it from another node. Run it on a cron or after a deploy, compare the Performance or SEO score against a threshold, and branch to an alert when a page regresses.

Hermes is an open agent based on Nous Research’s Hermes models. NoClick runs it for you and adds your connected apps as tools.

How do I give Hermes access to an app?

Connect the app's integration node to the agent's tools handle and pick the operations to expose. Each becomes a callable tool.

Do I need my own infrastructure?

No. The harness runs in a hosted sandbox and NoClick handles credentials for the apps you connect.

Can Hermes use multiple apps at once?

Yes. Wire in several integrations and the agent can use them together in a single run.

Connect Hermes to Google PageSpeed

Give an AI agent your Pagespeed node as a tool and it can measure a page's real performance itself instead of guessing at it. It can analyze a URL on mobile or desktop, pull the Performance, SEO, Accessibility, and Best Practices scores, read Core Web Vitals and lab metrics, and list the exact opportunities to fix, calling each one as a tool while it works. Wire it into a hosted coding agent in NoClick and it can audit a page, act on what it finds, and re-check the score afterward.

Loading preview…

Opens a workflow with 15 tools ready to use.

What happens next

Set up in minutes

The open button drops you into a guided setup. It asks exactly this, nothing else:

Choose where it runs
On your computer through your own Hermes, or hosted in the cloud. The rest of setup adapts to your choice:
Pick what it can do in Google PageSpeed
A preset of real Google PageSpeed operations. Each one becomes a tool the agent can call, and it can't touch anything outside the list.
Connect Google PageSpeed
One sign-in. The agent acts through your account, scoped to the operations you picked.
Add your personal MCP link
The last step mints a private MCP URL for this agent and walks you through exactly this, with your real link filled in. The page flips to Connected the moment your Hermes first calls in.
1. Run this in your terminal
$hermes mcp add noclick --url https://mcp.noclick.app/s/<your-link>
2. Answer no when asked if the server requires authentication — the link itself is the key
<your-link> is your personal server URL — setup mints it for you.

Exact capabilities

Tools Hermes gets

Pick the operations you want and each becomes a tool the agent can call directly while it works.

Google PageSpeed

15 tools

Run Analysis (Full)

Run a full Lighthouse + CrUX analysis and return the complete report.

Analyze (Mobile)

Run the analysis using the mobile crawler/emulation profile.

Analyze (Desktop)

Run the analysis using the desktop profile.

Performance Score

Run with the PERFORMANCE category and extract the 0-100 Lighthouse score.

Accessibility Score

Run with the ACCESSIBILITY category and extract the score + audits.

SEO Score

Run with the SEO category and extract the SEO score + failing audits.

Best Practices Score

Run with the BEST_PRACTICES category and extract the score + audits.

All Category Scores

Run all four Lighthouse categories in one call and extract every score.

Core Web Vitals (URL)

Extract real-user Core Web Vitals (CrUX field data) for the specific URL.

Origin Core Web Vitals

Extract aggregated CrUX field data for the entire origin (whole site).

Lab Metrics

Extract lab metrics (FCP, LCP, Speed Index, TBT, CLS, TTI) from Lighthouse.

Get Audit Details

Extract a single named Lighthouse audit (score + details).

Get Opportunities

Extract performance opportunity audits with estimated ms/byte savings.

Get Screenshot

Extract the final rendered screenshot (base64 data URI) of the page.

Localized Report

Run with a specific result locale for localized audit text.

Starting points

What you can build

Every morning, run a full analysis on your homepage and post the Performance score and Core Web Vitals to Slack so regressions get caught early.

When a deploy webhook fires, analyze the key landing pages on both mobile and desktop and flag any page whose Performance score drops below your threshold.

Pull the list of opportunities for a slow page, hand them to an AI step, and open a ticket with the ranked fixes.

On a weekly schedule, collect All Category Scores for a list of URLs and append them to a Google Sheet to track SEO and Accessibility over time.

Compare lab metrics against Origin Core Web Vitals for a URL to see whether real users experience the same speed your tests report.

Capture a screenshot and the Best Practices score for each new marketing page before launch and attach both to the review.

Meet the runner

About Hermes

Hermes is an open agent built on Nous Research’s Hermes models. NoClick runs it hosted and connects your apps to it as tools, so it can do real work across your systems. Wire an integration into the agent and Hermes can use that app’s operations directly while it reasons through a task. It is a strong pick when you want an open-model agent with genuine tool access.

Built on open Hermes models

Capable general reasoning

Flexible, open-model foundation

Hosted with apps wired in as tools

Good to know