Draft:Voxt

LLM-generated pages with certain obvious signs of being machine generated may be deleted without notice.

Draft:Voxt

Voxt
DevelopersActual Chat, Inc.
ReleaseMar 2, 2023; 3 years ago (Mar 2, 2023)
Written inC#, TypeScript
Operating systemWeb, Android, iOS, Windows
LicenseAGPLv3
Websitevoxt.ai
Repositorygithub.com/Actual-Chat/actual-chat

Voxt (formerly Actual Chat) is an open-source messaging application developed by Actual Chat, Inc. that combines real-time voice communication with automatic live transcription.[1] The application allows users to speak instead of typing, with their speech converted to text in real time, enabling recipients to either listen to the audio or read the transcript.[2] Voxt is available on the web, Android, iOS, and Windows.[3][4][5]

Overview

Voxt positions itself as a voice-first messenger that bridges the gap between text messaging and voice calls. The application's core premise is that speaking is approximately five times faster than typing, but traditional voice messages create an inconvenience for recipients who may not be in a position to listen. By automatically transcribing voice in real time, Voxt allows senders to communicate by voice while recipients can choose to read the text or play back the audio at their convenience.

Unlike social media platforms, Voxt does not feature public feeds, follower systems, or algorithmic content recommendations. The application is designed for private communication between known contacts.

Features

Voice and transcription

Voxt streams audio in real time using the Opus codec and transcribes speech as it is spoken. The live transcription appears to other participants while the speaker is still talking, similar to live captions. Multiple transcription engines are supported, including Deepgram, Google Cloud Speech-to-Text, and OpenAI's transcription models. The application also supports on-device transcription using NVIDIA's Parakeet TDT model via the ONNX Runtime, which processes speech locally without sending audio to cloud services.

Audio messages can be replayed after the fact, and the text transcript remains searchable.

AI features

Voxt integrates artificial intelligence capabilities powered by OpenAI models:

  • Conversation summaries — Automatic AI-generated summaries of conversations, allowing users to catch up on missed discussions.
  • Live translation — Real-time translation of messages with automatic language detection.
  • Semantic search — An AI-powered search system that uses machine learning embeddings via OpenSearch to find messages by meaning rather than exact keywords.
  • In-chat AI assistant — A built-in chatbot that can answer questions within conversations.

Communication modes

The application supports peer-to-peer private chats, group chats, and "Places" — workspace-style containers that can hold multiple chat rooms, similar in concept to Discord servers or Slack workspaces. Message threading, emoji reactions, file attachments (up to 500 MB), and message forwarding are also supported.

Privacy and security

Voxt is designed with an end-to-end encryption architecture using AES-256-GCM with ECDH P-256 key agreement and HKDF-SHA-256 key derivation. The application uses WebRTC audio processing for noise cancellation and echo suppression, and supports on-device voice activity detection using ONNX models.

Technology

Voxt is built on the .NET platform using C#. The web interface uses Blazor (a combination of server-side and WebAssembly rendering), while mobile and desktop applications are built with .NET MAUI sharing a Blazor WebView. The frontend also uses TypeScript and Tailwind CSS.

The backend uses PostgreSQL as its primary database, Redis for caching, and NATS for message brokering. Real-time state synchronization is handled by ActualLab.Fusion, an open-source framework developed by the same team that provides automatic computed-state invalidation and push updates. Communication between client and server uses SignalR WebSockets with MessagePack serialization.

In production, the application runs on Google Cloud Platform using Kubernetes, with AlloyDB (a PostgreSQL-compatible managed database) as the primary data store.

History

Voxt was founded in 2022 by Alex Yakunin and Alexey Kochetov as Actual Chat, incorporated as a Delaware corporation based in Laguna Niguel, California. The product was rebranded to Voxt with the domain voxt.ai, though the legal entity remains Actual Chat, Inc. The source code is publicly available on GitHub under the GNU Affero General Public License v3 (AGPLv3).

Platforms

Platform Availability
Web voxt.ai
Android @play.google.com (minimum Android 9)
iOS @apps.apple.com (minimum iOS 16.4)
Microsoft Windows @apps.microsoft.com (minimum Windows 10 version 1809)

See also

References

  1. ^ "Voxt: The New Voice of Everyday Communication". Voxt. Retrieved February 18, 2026.
  2. ^ Yakunin, Alex (March 2023). "Actual Chat is in Open Beta!". Medium. Retrieved February 18, 2026.
  3. ^ "Voxt – Apps on Google Play". Google Play. Google. Retrieved February 18, 2026.
  4. ^ "Voxt – Microsoft Store". Microsoft Store. Microsoft. Retrieved February 18, 2026.
  5. ^ "Actual Chat on the App Store". App Store. Apple. Retrieved February 18, 2026.


Category:Instant messaging clients Category:Free and open-source software Category:Software using the AGPL license Category:Cross-platform software Category:Speech recognition software Category:VoIP software Category:2022 software Category:Companies based in Orange County, California

Content Disclaimer

Informasi ini disarikan dari Wikipedia dan disajikan kembali untuk tujuan edukasi. Konten tersedia di bawah lisensi CC BY-SA 3.0. Kami tidak bertanggung jawab atas ketidakakuratan data yang bersumber dari kontribusi publik tersebut.

  1. The information displayed on this website is sourced in part or in whole from Wikipedia and has been adapted for the purpose of restating it. We strive to provide accurate and relevant information, however:
  2. There is no guarantee of absolute accuracy. Wikipedia is an open, collaborative project that can be edited by anyone, so information is subject to change.
  3. It is not intended to constitute professional advice. The content displayed is for informational and educational purposes only. For important decisions (e.g., medical, legal, or financial), please consult a professional.
  4. Content copyright. Wikipedia is licensed under the Creative Commons Attribution-ShareAlike License (CC BY-SA). This means that content may be reused with appropriate attribution and shared under a similar license.
  5. Responsible use. Any risk arising from the use of information from this website is entirely the responsibility of the user.