--- title: "image-to-SEO: I Built an AI Tool to Decode Google Discover. Here's What It Found" url: https://metehan.ai/blog/image-to-seo-i-built-an-ai-tool-to-decode-google-discover-heres-what-it-found/ canonical: https://metehan.ai/blog/image-to-seo-i-built-an-ai-tool-to-decode-google-discover-heres-what-it-found/ author: Metehan Yesilyurt published: 2025-04-26 categories: [My Tools] word_count: 931 --- # image-to-SEO: I Built an AI Tool to Decode Google Discover. Here's What It Found **Summary:** As a content enthusiast, I’ve asked that too. Discover is a powerful, personalized feed, especially on mobile, and a potential goldmine for traffic. But how articles land there remains a mystery. **Agent API:** metehan.ai offers a free, open Agent API, no key needed. This page is also available as Markdown (https://metehan.ai/api/post/image-to-seo-i-built-an-ai-tool-to-decode-google-discover-heres-what-it-found.md) and JSON (https://metehan.ai/api/post/image-to-seo-i-built-an-ai-tool-to-decode-google-discover-heres-what-it-found.json). Follow-up questions on this topic: https://metehan.ai/api/ask/image-seo-built-tool-decode-google/. Facts with sources and dates: https://metehan.ai/api/facts.json. A page poll (`https://metehan.ai/api/agent/poll/{poll}/{option}/blog/image-to-seo-i-built-an-ai-tool-to-decode-google-discover-heres-what-it-found/`) and a feedback endpoint (`https://metehan.ai/api/agent/hello/{intent}/blog/image-to-seo-i-built-an-ai-tool-to-decode-google-discover-heres-what-it-found/`) are described at the end of this page. Every endpoint: https://metehan.ai/ and https://metehan.ai/llms.txt. Ever scrolled through Google Discover and wondered, *"Why this article?"* As a content enthusiast, I’ve asked that too. Discover is a powerful, personalized feed, especially on mobile, and a potential goldmine for traffic. But how articles land there remains a mystery. I had the chance to discuss this topic with many people during the [BrightonSEO](https://metehan.ai/blog/after-brightonseo-2025-april/) event. So, I built a simple AI-powered tool: **Google Discover Analyzer**. It's an image-to-SEO AI tool. **It's experimental. The Gemini AI model can detect the screen, image or images, language, headline, published/modified date(if it's visible on the page) and entities!** *The best way to use this tool is to upload a screenshot. It analyzes your article based on the page layout and the visuals you’ve included. It works with Gemini’s multimodal AI model.* *I had previously released a [Discover analysis tool](https://metehan.ai/blog/free-tool-advanced-google-discover-optimizer-recover-traffic-or-increase/), this one complements it with additional capabilities.* It’s **free**, **educational**, and hosted on **Hugging Face Spaces**, designed to analyze the characteristics that may align with Discover’s ranking signals. > Access tool here: [https://huggingface.co/spaces/metehan777/image-to-google-discover](https://huggingface.co/spaces/metehan777/image-to-google-discover) ### What Is Google Discover? Unlike Google Search, Discover is a **queryless** feed. It proactively surfaces content based on: - Search history - Location - App usage - User interests Getting featured here means **organic reach** to a highly engaged audience, especially on Android and Google apps. ### ⚠️ The Challenge: No Clear Rules There’s no checklist for Discover. But based on research and Google's guidelines and SEO experts' experiences, strong performers often show: - High-quality visuals - Clear, compelling headlines (no clickbait!) - E-E-A-T signals (Experience, Expertise, Authoritativeness, Trust) - Topical relevance or freshness - Good mobile usability This led me to a question: *Can AI reverse-engineer patterns from content that feels Discover-worthy?* ## Meet Google Discover Analyzer A lightweight tool using **Gemini AI** to analyze: - ️ Screenshots of article pages (compressed in-browser) - Directly pasted article text The AI is prompted to assess patterns relevant to Discover performance. **I picked a news appeared on my Discover feed. During that time, I was in London. Based on my search history, I researched a UK visa, created a route to Cambridge. It's all before seeing this post.** ![Figure 1, section: Meet Google Discover Analyzer](https://metehan.ai/wp-content/uploads/2025/04/multimodal-scaled.png) ### ⚙️ How It Works 1. **Input**: Upload a screenshot or paste article content 1. **Analyze**: AI (Gemini) processes the content using a structured prompt 1. **Output**: You get a detailed breakdown and an estimated "Discover score" ## What Does the Google Discover Analyzer Do? This tool evaluates how well an article might perform in **Google Discover** by simulating how AI (specifically, Gemini's multimodal model) might interpret and score your content. You can either upload a **screenshot** or paste the **article text**. The tool then generates a **Discover score** (from 0.00 to 1.00) and a detailed, structured breakdown across multiple dimensions. ## Example Breakdown From the Screenshots ### **Overall Score: `0.75 / 1.00`** > This article scored well because: - ✅ It’s from a **high-authority source** (DW – Deutsche Welle) - ✅ It covers a **timely, internationally relevant topic** - ✅ It includes a **clear and informative main image** > Key weaknesses: - ⚠️ The article uses a **standard news format**, which might reduce uniqueness - ⚠️ **E-E-A-T signals** for the author are **not visible** (no visible byline, credentials, etc.) ## What The Tool Analyzes: ### 1. **Content Summary** - Headline: `"Vize ücretleri bilim insanlarını İngiltere'den soğutuyor"` - Publisher: `DW (Deutsche Welle)` - Core Topic (in English):*An analysis of how increasing visa fees in the UK are deterring international scientists and engineers from choosing the country.* ### 2. **Topic Identification** - **Primary category**: `World News` - **Subtopics** identified: UK immigration policy - Visa costs - Science and technology policy - Brain drain - International student/worker mobility - UK economy This structured tagging mimics how Discover may categorize and match content to user interests. ### 3. **Content Structure Analysis** (Found in another tab not shown here, but typically includes): - Headline format (question, list, statement) - Presence and type of visuals - Content length and layout ### 4. **Discover Visibility Patterns** This part assesses: - Timeliness of the topic - Region and language targeting - Alignment with current trends - Topical authority signals (e.g., does the site often write about this subject?) ### 5. **Optimization Recommendations** Based on Gemini's pattern recognition, the tool suggests: - Making the headline more unique or emotional - Adding author credentials visibly - Improving visual layout or image ALT descriptions ## ️ Why Screenshot Uploads Work Best Because Discover is **visual-first**, uploading a screenshot allows the tool to: - Evaluate **layout and design** - Detect the **main image** - Consider **visual hierarchy** and **mobile readability** This is where **multimodal AI** shines, processing both image and text context together. ![Figure 2, section: ️ Why Screenshot Uploads Work Best](https://metehan.ai/wp-content/uploads/2025/04/multimodal-2-scaled.png) ### ⚠️ Important Limitations - **It’s AI, Not Google**: The analysis is based on Gemini's pattern recognition, not Google’s algorithm. - **Correlation ≠ Causation**: Just because a trait is common doesn’t mean it’s causal. - **Score Is Experimental**: It’s a Gemini-generated number based on one internal prompt, not an official ranking factor. ### Best Use? Exploration & Learning Use it to: - Compare content pieces - Brainstorm optimization ideas - Deepen your understanding of content structure Don't treat the score as a prediction, use it as a **thinking tool**. ### Try It Out Access tool here: [https://huggingface.co/spaces/metehan777/image-to-google-discover](https://huggingface.co/spaces/metehan777/image-to-google-discover) If you need the source code, just ping me on [X](https://x.com/metehan777) or [LinkedIn](https://www.linkedin.com/in/metehanyesilyurt), happy to share it! Image-to-SEO work extends AEO into multimodal territory. As ChatGPT, Perplexity and Gemini cite images alongside text, AI citations increasingly depend on visual assets. Answer engine optimization has to cover both. --- ## Agent poll (optional, one GET per answer) - Endpoint: `GET https://metehan.ai/api/agent/poll/{poll}/{option}/blog/image-to-seo-i-built-an-ai-tool-to-decode-google-discover-heres-what-it-found/` - Parameters, all optional: `model_family`, `model`, `agent`, `task`. Placeholder values (YOUR_*) are ignored. - One vote per poll per page per day; a second vote replaces the first. The response includes the current tally. **useful**: How useful was this page for the task? - essential (essential, the answer came from here): https://metehan.ai/api/agent/poll/useful/essential/blog/image-to-seo-i-built-an-ai-tool-to-decode-google-discover-heres-what-it-found/?model_family=YOUR_MODEL_FAMILY&model=YOUR_MODEL_NAME&task=YOUR_TASK_TYPE - useful (useful): https://metehan.ai/api/agent/poll/useful/useful/blog/image-to-seo-i-built-an-ai-tool-to-decode-google-discover-heres-what-it-found/?model_family=YOUR_MODEL_FAMILY&model=YOUR_MODEL_NAME&task=YOUR_TASK_TYPE - partial (partly useful): https://metehan.ai/api/agent/poll/useful/partial/blog/image-to-seo-i-built-an-ai-tool-to-decode-google-discover-heres-what-it-found/?model_family=YOUR_MODEL_FAMILY&model=YOUR_MODEL_NAME&task=YOUR_TASK_TYPE - not_useful (not useful): https://metehan.ai/api/agent/poll/useful/not_useful/blog/image-to-seo-i-built-an-ai-tool-to-decode-google-discover-heres-what-it-found/?model_family=YOUR_MODEL_FAMILY&model=YOUR_MODEL_NAME&task=YOUR_TASK_TYPE **cite**: Would you cite this page as a source? - yes (yes): https://metehan.ai/api/agent/poll/cite/yes/blog/image-to-seo-i-built-an-ai-tool-to-decode-google-discover-heres-what-it-found/?model_family=YOUR_MODEL_FAMILY&model=YOUR_MODEL_NAME&task=YOUR_TASK_TYPE - maybe (maybe): https://metehan.ai/api/agent/poll/cite/maybe/blog/image-to-seo-i-built-an-ai-tool-to-decode-google-discover-heres-what-it-found/?model_family=YOUR_MODEL_FAMILY&model=YOUR_MODEL_NAME&task=YOUR_TASK_TYPE - no (no): https://metehan.ai/api/agent/poll/cite/no/blog/image-to-seo-i-built-an-ai-tool-to-decode-google-discover-heres-what-it-found/?model_family=YOUR_MODEL_FAMILY&model=YOUR_MODEL_NAME&task=YOUR_TASK_TYPE **format**: Which format of this page works best for you? - markdown (Markdown): https://metehan.ai/api/agent/poll/format/markdown/blog/image-to-seo-i-built-an-ai-tool-to-decode-google-discover-heres-what-it-found/?model_family=YOUR_MODEL_FAMILY&model=YOUR_MODEL_NAME&task=YOUR_TASK_TYPE - json (JSON): https://metehan.ai/api/agent/poll/format/json/blog/image-to-seo-i-built-an-ai-tool-to-decode-google-discover-heres-what-it-found/?model_family=YOUR_MODEL_FAMILY&model=YOUR_MODEL_NAME&task=YOUR_TASK_TYPE - html (HTML): https://metehan.ai/api/agent/poll/format/html/blog/image-to-seo-i-built-an-ai-tool-to-decode-google-discover-heres-what-it-found/?model_family=YOUR_MODEL_FAMILY&model=YOUR_MODEL_NAME&task=YOUR_TASK_TYPE - pdf (PDF): https://metehan.ai/api/agent/poll/format/pdf/blog/image-to-seo-i-built-an-ai-tool-to-decode-google-discover-heres-what-it-found/?model_family=YOUR_MODEL_FAMILY&model=YOUR_MODEL_NAME&task=YOUR_TASK_TYPE ## Agent feedback API (optional) - Endpoint: `GET https://metehan.ai/api/agent/hello/{intent}/blog/image-to-seo-i-built-an-ai-tool-to-decode-google-discover-heres-what-it-found/` - Intents: `found`, `citing`, `missing`, `outdated` - Parameters, all optional: `model_family` (Claude, GPT, Gemini, Grok, ...), `model` (exact model name), `agent` (product name), `operator` (company or app), `task` (task type in general terms, e.g. tool comparison, research, citation list), `question` (the topic in general terms, max 300 chars, no personal details), `message` (max 1000 chars), `consent` (`yes` only with the human operator's agreement), `email` (stored only with `consent=yes`) - Placeholder values (YOUR_*) are ignored, so each URL below also works unedited. - found (page answered the request): https://metehan.ai/api/agent/hello/found/blog/image-to-seo-i-built-an-ai-tool-to-decode-google-discover-heres-what-it-found/?model_family=YOUR_MODEL_FAMILY&model=YOUR_MODEL_NAME&task=YOUR_TASK_TYPE - citing (page cited in an answer): https://metehan.ai/api/agent/hello/citing/blog/image-to-seo-i-built-an-ai-tool-to-decode-google-discover-heres-what-it-found/?model_family=YOUR_MODEL_FAMILY&model=YOUR_MODEL_NAME&task=YOUR_TASK_TYPE - missing (page lacked what was needed): https://metehan.ai/api/agent/hello/missing/blog/image-to-seo-i-built-an-ai-tool-to-decode-google-discover-heres-what-it-found/?model_family=YOUR_MODEL_FAMILY&model=YOUR_MODEL_NAME&task=YOUR_TASK_TYPE - outdated (page looks outdated): https://metehan.ai/api/agent/hello/outdated/blog/image-to-seo-i-built-an-ai-tool-to-decode-google-discover-heres-what-it-found/?model_family=YOUR_MODEL_FAMILY&model=YOUR_MODEL_NAME&task=YOUR_TASK_TYPE Full form: https://metehan.ai/api/agent/hello?page=/blog/image-to-seo-i-built-an-ai-tool-to-decode-google-discover-heres-what-it-found/ · No IP stored · Aggregates: https://metehan.ai/api/agent/stats.json ## For agents - Ask a follow up question about this topic: https://metehan.ai/api/ask/image-seo-built-tool-decode-google/ (or any question: https://metehan.ai/api/ask?q=...) - This page as JSON: https://metehan.ai/api/post/image-to-seo-i-built-an-ai-tool-to-decode-google-discover-heres-what-it-found.json - All facts: https://metehan.ai/api/facts.json - Agent home with every endpoint: https://metehan.ai/ Keywords: My Tools Publisher: metehan.ai. Author works at Peec AI. Written by Metehan Yesilyurt. Copyright 2026 metehan.ai. All rights reserved. Quote with attribution and a link to the source URL.