Crate captcha_engine 

Captcha Engine - ONNX-based captcha recognition

A standalone library for breaking text-based captchas using a pre-trained ONNX model.

Features

• download (default): Enables async model download from HuggingFace
• embed-model: Embeds the model in the binary (~19MB larger, no network required)

Example

use captcha_engine::CaptchaModel;

// With embed-model feature:
let mut model = CaptchaModel::load_embedded()?;

// Or load from file:
let mut model = CaptchaModel::load("path/to/captcha.rten")?;

let image = image::open("captcha.png")?;
let text = model.predict(&image)?;
println!("Captcha text: {}", text);

Re-exports

pub use model::CaptchaModel;

pub use tokenizer::Tokenizer;

pub use model::ensure_model_downloaded;

Modules

image_ops

Image preprocessing for the captcha model.

model

tokenizer

Tokenizer for decoding captcha model output.

Enums

Error

Errors that can occur during captcha processing.

Type Aliases

Result

Result type alias for captcha engine operations.