漢字 Chinese Character Lookup — 解字
Upload a photo of a Chinese character to 解字 (jiězì) — identify it and look it up in three dictionaries (English-translated), in the spirit of 說文解字 (Shuōwén Jiězì).
Built for friends who can't yet type, write, or pronounce Chinese — but can snap a photo. Excellent dictionaries already explain where each character comes from; this tool is just the missing entry point, turning a photo into a character they can look up.
Click to choose a photo or drag & drop it here
Works best with a clear, cropped photo of the character
Best experience: open Original 原文 and use your browser's built-in translation (right-click → "Translate to English") — you get English text and all images. The "English (translated)" links use Google's proxy, where some ancient-script images may not display.
解字 (jiězì) means, literally, to "unlock" a character — to see the picture or idea it started from, thousands of years ago, when writing was first invented. 說文解字 (Shuōwén Jiězì), compiled by the scholar Xǔ Shèn around 100 CE, was China's first systematic dictionary of these origins, and it's where this project takes its name.
Several excellent modern databases carry that work forward today — CUHK's Multi-function Chinese Character Database, zi.tools, and Taiwan's Dictionary of Chinese Character Variants among them. But they all assume you can already type the character, know how to pronounce it, or can describe its strokes. If you can't do any of that yet — a young learner, a heritage speaker, a total beginner — you're locked out of resources that would otherwise help you the most.
This tool doesn't try to replace those dictionaries. It's the missing on-ramp to them: take a photo, get the character, and go explore.
Traditional scholarship — first codified in 說文解字 — groups the way characters were formed into six categories:
- 象形 (xiàngxíng) — pictographs, literal pictures of things: 人 (a person, seen from the side), 木 (a tree, with branches and roots), 山 (mountain peaks).
- 指事 (zhǐshì) — indicators, abstract symbols marking an idea or position: 上 ("above") and 下 ("below") mark a position relative to a line.
- 會意 (huìyì) — compound ideographs, meaning built by combining pictures: 休 combines 人 (person) and 木 (tree) — a person resting against a tree, meaning "to rest."
- 形聲 (xíngshēng) — phono-semantic compounds, by far the most common type today, pairing a meaning component with a sound component.
- 轉注 (zhuǎnzhù) — mutually explanatory characters, related in both meaning and sound.
- 假借 (jiǎjiè) — phonetic loans, where a character is borrowed for its sound rather than its original meaning.
Knowing which category a character falls into is usually the fastest way to understand — and remember — where it comes from.
象形: a pictograph of a person seen from the side, standing on two legs.
象形: a pictograph of a tree, with a trunk, branches reaching up, and roots reaching down.
會意: literally 人 next to 木 — a person leaning against a tree to rest.
Everything runs in your browser. Tesseract.js, an open-source OCR engine, reads the photo after several image-cleanup passes — auto-cropping to the ink, converting to black-and-white with adaptive thresholding, and normalizing the size — then ranks the most likely characters by confidence. Your photo is never uploaded to a server; recognition happens entirely on your device.
Once you confirm the character, links open directly into three character-origin dictionaries, each in both its original language and an English-translated version, so you can dig into etymology, historical forms, and readings straight away.