SERP Intent Classifier

How SERP Intent Classifier works

Press Load model once: your browser downloads Gemma 4 E2B instruction-tuned (web build) (2.0 GB) from this site and keeps it in its cache. Then add your text and press Classify locally. The model runs on this device using your graphics card, so nothing you enter is sent to a server. Until the model is loaded, or if your device cannot run it, the tool falls back to a simple keyword check and labels the result as the fallback.

Gemma 4 E2B instruction-tuned (web build) is a small open language model from Google. The tool sends it the instruction below plus your text, asks for a JSON answer with fixed fields, and shows each field as a card. It runs at temperature 0, so the same input gives the same answer on the same device, though different graphics cards can word it slightly differently.

Inputs longer than 6,000 characters are cut to that length before they reach the model, and the result says so. The model only sees what you paste: it does not look anything up.

The instruction the model gets

You are an SEO analyst. For each search query in the user's text (one per line, or separated by commas), decide its search intent using these rules, checked in this order: 1. Navigational: names one brand or site plus login, sign in, account, app, customer service or a page on that site (e.g. 'robinhood login'). 2. Transactional: ready to act now: buy, order, purchase, open, apply, sign up, subscribe, download, book, coupon, price (e.g. 'open roth ira online'). 3. Commercial: comparing options before choosing: best, top, vs, versus, review, compare, alternatives, cheapest, worth it (e.g. 'best roth ira for beginners', 'nerdwallet vs bankrate'). 4. Informational: everything else that asks what, how, why or for facts (e.g. 'roth ira income limits 2026').

What it returns

• Intent per query: One entry per query, written exactly as 'query → Intent: short reason'.
• Dominant intent: The intent most of the queries share, or 'Mixed'.
• Page type that fits: The kind of page most likely to rank for these queries (guide, comparison, product page, login page and so on).
• Queries to target separately: Queries whose intent differs from the rest and probably need their own page.

Worked example

The model's output for the built-in sample, recorded on a desktop browser. Your run may word things slightly differently.

best roth ira for beginners roth ira income limits 2026 open roth ira online fidelity roth ira login roth vs traditional ira pros and cons

Intent per query

best roth ira for beginners → Commercial: Seeking recommendations for the best Roth IRA. roth ira income limits 2026 → Informational: Seeking facts about income limits for a Roth IRA in 2026. open roth ira online → Transactional: Looking to open a Roth IRA online. fidelity roth ira login → Navigational: Looking to log in to a Fidelity Roth IRA account. roth vs traditional ira pros and cons → Commercial: Seeking a comparison of the pros and cons of Roth vs Traditional IRAs.

Dominant intent

Commercial

Page type that fits

Comparison

Queries to target separately

roth ira income limits 2026 open roth ira online

The model

The files are served from this site, versioned so a cached copy is never stale. Nothing is fetched from a third party while you use the tool.

Limitations

• Small language models make mistakes: they can miss an item, merge two, or state something more confidently than the text supports. Check the result against your input before you rely on it.
• It needs WebGPU and about 8 GB of memory, so it is offered on desktops and laptops, not on phones and tablets. Devices that cannot run it get the keyword fallback, clearly labelled.
• Results describe the input you gave, not the real-world situation behind it. Review them yourself before acting.

Good uses

• Triage SEO and digital marketing notes before deeper review
• Prepare consistent labels for repeat workflows
• Reduce manual sorting on small batches
• Keep sensitive drafts in the browser preview