One link, one card set
AI CONTENT PRESS
Paste a link or topic —…
59Auto cover & hashtags
59 layouts × 75 palettes
Paste a link or topic — AI breaks it into "one p
Export 2x PNG and publish.
AI CONTENT PRESS · 图文内容工作室
Paste any link or topic. ToCard breaks it into a series of sharp, fully editable cards — each making a single point — with a cover and hashtags, ready for Xiaohongshu, LinkedIn or anywhere.
Scene presets
AI CONTENT PRESS
Paste a link or topic —…
59Auto cover & hashtags
59 layouts × 75 palettes
Paste a link or topic — AI breaks it into "one p
Export 2x PNG and publish.
HOW THE PRESS WORKS
Paste a link, a topic or raw text. The press fetches and reads it.
AI splits the content into a one-point-per-card series with cover and hashtags.
Edit any word, switch ink and paper, export 2x PNG sets and post.
FAQ
ToCard is an AI content studio for social publishing. Paste a link or a topic, and it reads the source, extracts the key ideas and presses them into a set of knowledge cards — one point per card, each with a cover, hashtags and a structured summary. The whole flow takes about 30 seconds, and every card exports as a publish-ready 2x PNG.
Since October 2026 generation runs entirely on your own API key — ToCard keeps no server-side model credentials, so quota and cost stay under your control. Open "Model settings" in the top-right, pick a provider (DeepSeek or any OpenAI-compatible endpoint), fill in the model name and key, and you are set. The key is saved only in your browser's localStorage and sent with each generation request — it is never stored in our database. Without a key, generation is unavailable by design.
Some sites block automated fetching or render their pages with JavaScript, so only an empty shell can be extracted (Zhihu and other TLS-fingerprint-protected sites are typical cases). When that happens ToCard tells you explicitly instead of producing meaningless text. The most reliable fix: copy the article text and paste it into the topic input — the quality is the same, often better.
Yes — every word on every card is editable, both inline on the card itself and through the side editor. You can switch between 59 layouts and 75 ink & paper palettes at any time without losing your content, re-select which cards stay in the set, and export only when it looks right.
Each card exports as a 2x PNG at 1080×1440 (3:4), one by one or as the whole set bundled in a .zip. A visual render mode can also turn an entire card into a single AI-generated image. Sizes are tuned for Xiaohongshu, LinkedIn and WeChat — no cropping needed.
Yes. The images you export are yours to publish, including commercial use. That said, you are responsible for the source material you paste in and for how the generated content is used — make sure you have the rights to the original content.
The site itself runs in 7 languages: English, Chinese, French, Japanese, Russian, Arabic and Spanish. Independently of the interface, you can choose the output language for the generated cards — a Chinese article can become English cards and vice versa.
Your API key lives only in your browser and is never stored on our servers. Pasted text and fetched pages are used for the current generation only. If you log in and explicitly save a card set, it is kept in your account so you can come back to it; otherwise nothing is retained.
Free mini tools