cn2an

Why we included this project

Chinese-language products have a recurring data problem: numbers arrive as characters, not digits. cn2an converts between the two in both directions, and it handles the awkward cases that trip up naive parsers, including uppercase financial characters, mixed forms like "1百23", decimals, negatives, and dates, fractions, percentages, and temperatures embedded in sentences. That makes it useful for ASR post-processing, where a recognizer returns "一百二十三" and downstream logic needs 123, and for normalizing text before it reaches analytics. The API is small enough to wire into a pipeline in minutes, and an HTTP endpoint is available if you would rather call it as a service。