Upload any image. Add context. Watch Google Gemini craft three hilarious, sarcastic meme variants — ready to edit and download in seconds.

MemeGemini is a browser-based meme creation tool powered by Google Gemini's multimodal AI. It analyzes the visual content of any image you upload — reading expressions, body language, energy, and context — and automatically generates 3 distinct, caption-ready meme variants with intelligent placement, dynamic font sizing, and real-time canvas editing.

No accounts. No servers. Just your image, your Gemini API key, and a lot of sarcasm.

Memes are one of the highest-engagement content formats on the internet — but creating ones that actually land takes time, creativity, and a pulse on current culture. Memegemini removes that bottleneck.

Digital Marketers Stop spending hours brainstorming captions. Upload a campaign visual, describe your angle, and get 3 ready-to-post meme variants in seconds. Test different tones across variants without any extra effort. Meme campaigns deliver ~19% click-through rates compared to ~6% for standard marketing campaigns — and 14% higher CTR than email marketing.

Brands & Social Media Teams Meme content drives organic reach that polished ad creatives rarely match. Memes generate 10x more organic reach and 60% higher engagement than regular marketing visuals. Over 60% of consumers say they are more likely to purchase from brands that use memes — and 80% find such brands more relatable. Memegemini lets brand teams produce on-trend, humorous content quickly — without needing a dedicated creative team or an agency.

Film & Entertainment Marketing Memes have become a core pillar of modern film promotion. Meme-driven campaigns consistently deliver 18–22% engagement rates and hundreds of millions of impressions — often at a fraction of traditional ad spend. Turn posters, stills, or scene grabs into shareable content that builds organic buzz before a single ticket is sold.

Content Creators & Influencers Turn any image into shareable content quickly to stay consistent

• Visit Google AI Studio and generate a free API key.
• In the app, click "Configure API Key" in the top-right header and paste your key.
• Your key is stored only in your browser's local storage — it is never sent to any third-party server.

• Drag and drop any image onto the upload zone, or click "Browse File" to select one.
• Supported formats: PNG, JPG, JPEG, WEBP.

• Type a custom theme or topic in the text box — e.g., "developer debugging production on a Friday evening" or "cat trying to understand taxes".
• If left blank, Gemini defaults to universally relatable, everyday humor.

• Select your preferred Gemini model from the dropdown:
  • gemini-3.1-pro-preview (Recommended)
  • gemini-2.5-pro
  • gemini-2.5-flash

• Hit "Generate 3 Meme Variants" and watch the progress bar as Gemini analyzes your image and composes the captions.

• Each of the 3 output variants has its own live editing panel:
  • Customize top and bottom captions
  • Adjust font size, font family, text color, outline color, and outline thickness
  • Changes are reflected on the canvas in real-time
• Click "Download Meme" to save a variant as a PNG, "Copy" to copy it directly to your clipboard, or "Download All" to grab all three at once.

• Click "Try with Demo Mode" to load pre-built mock captions and explore the editor without needing a key.

AI-Powered Visual Analysis Gemini reads the uploaded image at a multimodal level — identifying characters, expressions, visual energy, and situational context — to produce captions that are genuinely relevant to the image, not generic placeholder text.

The Law of Emotional Inversion At the core of the caption engine is a humor formula: high-energy images get low-stakes, mundane captions; calm images get dramatic, existential reactions. This contrast is what makes memes actually funny.

3 Distinct Humor Variants Per Generation Every generation produces three stylistically different takes:

• Classical internet sarcasm
• Modern meta / self-deprecating humor
• Creative situational comedy

Dynamic Caption Placement The AI decides whether a meme works best with top text only, bottom text only, or both — keeping the layout clean and uncluttered per variant.

AI-Suggested Font Sizing Caption font size is dynamically scaled by the AI based on caption length (short punchy captions get large text; longer captions get smaller, fitted text) and further adjusted to match the native resolution of your image.

Real-Time Canvas Editing Every styling control — text content, font, size, color, outline — updates the meme canvas live as you type or adjust, with no re-generation needed.

Smart Word-Wrap Engine A built-in canvas compositing engine handles multi-line text wrapping, drop shadows, and stroke rendering natively to produce crisp, properly formatted output at any image resolution.

Custom Topic Support Provide a specific theme and Gemini will align all 3 variants to that context while still applying emotional inversion and visual analysis — making it useful for niche communities, professional inside jokes, or themed content.

Privacy-First API Key Handling Your Gemini API key is stored locally in the browser. The app makes calls directly from your browser to Google's API — no backend, no key exposure to any intermediary.

Demo Mode No key? No problem. A built-in demo mode loads pre-packaged mock captions so you can explore the full editing experience immediately.

• Vanilla HTML / CSS / JavaScript — zero frameworks, zero build steps
• Google Gemini API (/v1beta/models/:model/generateContent) with structured JSON schema response
• HTML5 Canvas API for meme compositing and text rendering
• Browser LocalStorage for API key persistence

No installation, no build process, no terminal required.

For non-technical users:

1. On this GitHub page, click the green "Code" button near the top right.
2. Select "Download ZIP" from the dropdown.
3. Once downloaded, find the ZIP file (usually in your Downloads folder) and extract/unzip it.
4. Open the extracted folder and double-click index.html.
5. It will open directly in your browser — no installation needed.

For developers:

git clone https://github.com/your-username/memegemini.git
cd memegemini
open index.html

Or serve it with any static file server:

npx serve .

🔍 Google Search Grounding (Planned) The next major feature will enable Google Search Grounding via the Gemini API — allowing the model to pull live, real-time information from the web as part of the generation call itself. No separate search tool or pipeline needed; Gemini will natively query Google Search before composing captions, grounding the humor in current trending topics, viral moments, and breaking news. The result: memes that are sharper, more timely, and far more likely to actually land.

All captions and jokes are dynamically generated by AI and do not reflect the views of the tool creator or the AI labs. Use responsibly. Note: You can update System instructions to fine-tune the tool for specific use cases or content policies.

Note: This is an independent personal project, not affiliated with or endorsed by Google. All trademarks belong to their respective owners.