moshi-chat

A language learning app that practices everyday conversations through role-playing and provides feedback on pronunciation and grammar.

Online text-to-speech service that converts text into multilingual MP3 audio with adjustable voice parameters

An English pronunciation training app that delivers personalized feedback and practice paths based on your native language.

Multilingual translation platform supporting real-time translation of text, images, and audio.

A language learning platform that creates annotated flashcards from authentic content using spaced repetition.

A non-fiction book summary platform offering distilled core content in text and audio formats.

SoundHound AI is a leading voice AI platform that empowers brands and enterprises to build custom voice-powered products, services, and applications. The platform combines proprietary speech-to-meaning and deep natural language understanding technologies with integrated third-party generative AI models such as ChatGPT. It enables real-time, context-aware conversations, allowing users to interact naturally through voice to handle complex queries across multiple domains, including weather, music, navigation, and customer service. SoundHound AI also powers the popular SoundHound Music app, which can identify songs from humming or singing, and includes AI components such as speech recognition, text-to-speech, and custom brand wake words. The solution is scalable, supports multiple languages, and is deployable on both edge and cloud environments, delivering personalized voice experiences and valuable data insights for brands.

David AI focuses on building the foundational data layer for audio AI by collecting, generating, and annotating large-scale, studio-grade audio datasets. Its proprietary dataset comprises over 10,000 hours of high-fidelity, speaker-separated audio spanning more than 15 languages, with rich metadata covering diverse accents and dialects. This diverse dataset empowers leading AI labs and companies to develop more natural, robust, and reasoning-capable advanced speech models. The David AI platform enables exponential scaling of audio data collection, effectively addressing the scarcity and fragmentation of high-quality audio data in the AI industry.

Vocode is an open-source library and hosted service for developers, designed to simplify the development of voice AI applications based on large language models (LLMs). By integrating speech recognition, natural language understanding, and text-to-speech synthesis, Vocode enables real-time conversational orchestration, empowering developers to build interactive voice agents for use cases such as phone calls, Zoom meetings, and personal assistants. Vocode supports multilingual capabilities and seamlessly integrates with leading voice and AI service providers, offering flexibility and scalability for diverse voice AI applications.

Truecaller is a trusted communication app that identifies unknown callers, blocks spam calls and messages, and delivers a secure, organized calling and messaging experience. Leveraging a global database of over 450 million users and AI-powered analysis, it helps users avoid scams, telemarketing, and robocalls. Featuring advanced capabilities such as AI call scanning, smart SMS categorization, and an AI-powered call screening assistant, Truecaller serves as a comprehensive solution for managing personal and business communications.

Elsa Speak is an advanced AI-powered language learning platform designed to help non-native English speakers improve their pronunciation, fluency, and overall speaking skills. Leveraging cutting-edge speech recognition technology, Elsa provides instant, detailed feedback on pronunciation, intonation, and rhythm. The platform offers personalized lessons, real-life scenario role-plays, and accent training, enabling learners at all levels to speak English with confidence and naturalness.

Friend is an innovative AI-powered wearable device designed as a pendant that serves as a personal companion to help users combat loneliness and provide encouragement in daily life. It listens to conversations via Bluetooth, processes the audio content, and sends supportive messages to the user's smartphone via text. Priced at $99 with no subscription fee, Friend prioritizes privacy by not storing audio or text data long-term, and all data is encrypted. Currently available in the United States and Canada, it supports iOS devices with Android compatibility coming in the future. The device is designed to create a genuine, judgment-free interactive experience, offering warm companionship rather than functioning as a productivity tool.