Changelog

New

• MakeMidas CLI: npm i -g @makemidas/cli, then makemidas image "…" or makemidas video "…". Coding agents like Claude Code, Cursor and Codex can run it too.
• Agent skill for Claude Code, Cursor and other agents: npx skills add makemidas-ai/skills.
• Setup guides for ChatGPT, Claude and Notion, with prompts to try.

Improved

• When the sign-in check can't verify your browser, the sign-in, sign-up and reset forms now say so and offer Try again, instead of leaving the button greyed out.

New

• MCP server at https://makemidas.com/api/mcp: make images and videos from ChatGPT, Claude, Claude Code, Cursor and other MCP apps with your credits. In ChatGPT and Claude, results appear in a gallery in the chat.
• REST API and API keys: create a key in Settings → API Keys and call /api/v1. The OpenAPI spec is at /openapi.json.
• Save to Notion: send any result to your Notion workspace as a page with the picture or video, the prompt and the settings.
• Settings → Connections: see the apps you've connected and disconnect any of them at once.
• Developers page with setup steps for each app.

Privacy

• Pet portraits stay private: they're never shown in the public gallery unless you agree, and never used to train AI models.

Improved

• The generator opens on the cheapest image and video models and the shortest clip length, so you only pay more when you choose to.

New

• New models: Flux 3 Klein, Seedance 2.0 and Seedance 2.0 Mini, Veo 3.1 Fast and Wan 3.0 Prime.
• Several reference pictures for Seedance videos.
• Full privacy on Ultimate and Max: your results are never shown in the public gallery or used to train AI models.

Improved

• The model picker is grouped by maker, with a sample result for each model.
• The aspect ratio follows the shape of your first picture.
• New example images for every image model.

Policy

• Results made without an Ultimate or Max plan may be shown in the public gallery without your name and may be used to train AI models. Pictures you upload and your prompts are never shown or used for training.

Improved

• Real results on every tool page, made with the models you can pick.
• New video examples from Veo, Kling, Wan and Grok Imagine.

Billing

• Checkout now calculates sales tax or VAT for your location. Listed prices don't include tax.

New

• Image and video generator: describe what you want, or start from a picture, and make images or video clips with leading AI models. It's now the home page.
• Tool pages for text to image, image to image, text to video and image to video.
• AI enhance rewrites a short idea into a detailed prompt, and style presets set the look in one click.
• Reuse a result as a reference picture, as the first frame of a video, or with the same settings.
• Drag and drop or paste pictures into the generator.

Plans and credits

• Everything runs on credits. Pro, Ultimate and Max plans add credits every month, and credit packs top up at any time.
• Every signed-in account gets free daily credits for the free models (free results are watermarked).
• Pet portraits now use credits too.

New

• AI pet portraits: upload a photo of your pet and get a royal portrait, a cartoon, a photo of you with your pet, or your pet added to a family photo.
• Free preview: see a watermarked preview before you buy, then download the full-resolution portrait.
• Sign in with Google, or with your email and a password.
• Help and policies in English and Chinese.

Privacy

• Photos you upload are deleted within 48 hours.