Open Source C Artificial Intelligence Software for Mobile Operating Systems

whisper.cpp is a lightweight, C/C++ reimplementation of OpenAI’s Whisper automatic speech recognition (ASR) model—designed for efficient, standalone transcription without external dependencies. The entire high-level implementation of the model is contained in whisper.h and whisper.cpp. The rest of the code is part of the ggml machine learning library. The command downloads the base.en model converted to custom ggml format and runs the inference on all .wav samples in the folder samples. whisper.cpp supports integer quantization of the Whisper ggml models. Quantized models require less memory and disk space and depending on the hardware can be processed more efficiently.

Presage (formerly Soothsayer) is an intelligent predictive text entry system. Presage generates predictions by modelling natural language as a combination of redundant information sources. Presage computes probabilities for words which are most likely to be entered next by merging predictions generated by the different predictive algorithms. Presage's modular and extensible architecture allows its language model to be extended and customized to utilize statistical, syntactic, and semantic predictive algorithms. Presage's predictive capabilities are implemented by predictive plugins. Predictive plugins use services provided by the platform to implement multiple prediction techniques.

shadowhook is an open-source native code hooking library for Android — designed to let developers intercept and override native (C/C++) functions inside Android apps at runtime. It supports both ARM32 and ARM64 architectures (and the respective “thumb” instruction set) and works across a wide range of Android OS versions. The library allows you to specify hook targets either by function address or by library name + function name, and it automatically handles newly loaded shared libraries (ELFs), ensuring hooks remain effective even when code is dynamically loaded at runtime. ShadowHook strives to be stable and safe for production use: it automatically avoids circular-call recursion between original and proxy functions, and supports unwind/unwind-safe proxy functions (CFI/FP unwind) so that backtraces, debugging, and stack unwinding remain correct. It also offers optional intercept and logging capabilities, letting developers record hooked calls.

Run a fast ChatGPT-like model locally on your device. This combines the LLaMA foundation model with an open reproduction of Stanford Alpaca a fine-tuning of the base model to obey instructions (akin to the RLHF used to train ChatGPT) and a set of modifications to llama.cpp to add a chat interface. Download the zip file corresponding to your operating system from the latest release. The weights are based on the published fine-tunes from alpaca-lora, converted back into a PyTorch checkpoint with a modified script and then quantized with llama.cpp the regular way.

MozJPEG improves JPEG compression efficiency achieving higher visual quality and smaller file sizes at the same time. It is compatible with the JPEG standard, and the vast majority of the world's deployed JPEG decoders. MozJPEG is compatible with the libjpeg API and ABI. It is intended to be a drop-in replacement for libjpeg. MozJPEG is a strict superset of libjpeg-turbo's functionality. All MozJPEG's improvements can be disabled at run time, and in that case it behaves exactly like libjpeg-turbo. MozJPEG is meant to be used as a library in graphics programs and image processing tools. We include a demo cjpeg command-line tool, but it's not intended for serious use. We encourage authors of graphics programs to use libjpeg's C API and link with MozJPEG library instead. Progressive encoding with "jpegrescan" optimization. It can be applied to any JPEG file (with jpegtran) to losslessly reduce file size.

Fast Artificial Neural Network Library is a free open source neural network library, which implements multilayer artificial neural networks in C with support for both fully connected and sparsely connected networks. Cross-platform execution in both fixed and floating point are supported. It includes a framework for easy handling of training data sets. It is easy to use, versatile, well documented, and fast. Bindings to more than 15 programming languages are available. An easy to read introduction article and a reference manual accompanies the library with examples and recommendations on how to use the library. Several graphical user interfaces are also available for the library.

Use Tesseract OCR in iOS 9.0+ projects written in either Objective-C or Swift. Easy and fast. Tesseract OCR iOS and TesseractOCR.framework are distributed under the MIT license. These are the current versions of the upstream bundled libraries within the framework that this repository provides, Tesseract 3.03-rc1, Leptonica 1.72, Image libraries, Libtiff 4.0.4, Libpng 1.6.18, and Libjpeg 9a.

Cactus is a low-latency, energy-efficient AI inference framework designed specifically for mobile devices and wearables, enabling advanced machine learning capabilities directly on-device. It provides a full-stack architecture composed of an inference engine, a computation graph system, and highly optimized hardware kernels tailored for ARM-based processors. Cactus emphasizes efficient memory usage through techniques such as zero-copy computation graphs and quantized model formats, allowing large models to run within the constraints of mobile hardware. It supports a wide range of AI tasks including text generation, speech-to-text, vision processing, and retrieval-augmented workflows through a unified API interface. A notable feature of Cactus is its hybrid execution model, which can dynamically route tasks between on-device processing and cloud services when additional compute is required.

*** This page is no longer used by the Dasher project - please see http://www.inference.phy.cam.ac.uk/dasher/ *** Dasher is a data entry interface incorporating language modelling and driven by continuous two-dimensional gestures, e.g. a mouse, a stylus

LLMFarm is a framework designed to simplify the deployment, management, and utilization of large language models in local or self-hosted environments, focusing on accessibility and efficient resource usage. It enables users to run LLMs on personal hardware or private infrastructure, reducing dependency on external APIs and improving data privacy. The system typically provides a user-friendly interface for loading models, configuring inference parameters, and interacting with them through chat or task-based workflows. It emphasizes modularity, allowing users to integrate different models, backends, or tools depending on their needs and hardware capabilities. LLMFarm is particularly useful for developers and researchers experimenting with local AI systems, as it lowers the barrier to entry for running and testing models without extensive setup. It also supports optimization techniques to improve performance on limited hardware, making it viable for smaller-scale deployments.

memory-leak-detector is a native memory-leak monitoring tool developed by ByteDance (originally for their video-app team) that helps developers detect, diagnose, and manage memory leaks in native (e.g. C/C++) code across Android applications. It offers relatively simple integration and wide monitoring coverage: by configuring the tool (or via broadcast control), developers can start leak detection for specific native libraries or entire processes. Upon activation, it tracks allocations / memory maps, logs usage, and outputs reports that can be post-processed (e.g. via provided Python scripts) to list potential leak suspects — assisting in identifying leaks that traditional tools might miss. The detector supports stack unwinding, mapping analysis, and reporting, making it suitable for complex applications where native memory misuse can lead to crashes or performance degradation. It’s been used within major ByteDance apps for native-memory leak governance.

Simulates a biologically diverse tropical island, and the ape inhabitants cognitive processes. For MacOS X, iPad, Windows and Linux. Features a non-polygonal graphics engine and a command-line version

OpStation_Qt: Operator Station for controlling multiple robots/robotics/devices using Player/Stage/Gazebo written using the open Qt GUI libraries. Written in C and C++. A good platform for storing data from scientific experiments as well as recreation

Scalable Language API (SLAPI) The most comprehensive architecture for conversational natural-language applications including speech recognition/synthesis, semantics, & machine translation. Used on Android & other mobile app platforms.