Skip to content

Project Timeline & History

Entries marked ??? (or with a partial year like 202?) have dates that could not be confirmed from a public source; they are left explicit rather than guessed.

In a nutshell

OVOS did not appear out of nowhere; it grew out of an earlier voice assistant called Mycroft, and other projects branched off along the way. This page is the family tree showing where everything came from and how the pieces relate. It is background reading for anyone curious about the project's history. See the Glossary.

Family Tree

Fork tree from mycroft-core: one line goes through mycroft-lib to ovos-core (marked "you are here"), a separate mycroft-dinkum branch that was abandoned, and a neon-core branch that was recreated on mycroft-lib before later moving to ovos-core

  • mycroft-core created
  • neon-core forked from mycroft-core
  • chatterbox forked from mycroft-core (closed source)
  • mycroft-lib forked from mycroft-core to become a library, it is only a properly packaged mycroft-core
  • chatterbox rewritten to use mycroft-lib internally, no longer a hard fork
  • neon-core rewritten to use mycroft-lib internally, no longer a hard fork
  • mycroft-lib renamed to HolmesIV to avoid trademark issues
  • HolmesV created to use HolmesIV internally, HolmesV gets features and HolmesIV bug fixes only
  • chatterbox updated to HolmesV
  • neon-core updated to HolmesV
  • HolmesIV+HolmesV abandoned, chatterbox focus on closed source product
  • ovos-core forked from HolmesV
  • neon-core updated to ovos-core
  • mycroft-dinkum forked from mycroft-core

Events timeline

2015

2016

  • Feb 2016 - MycroftAI Mimic TTS released
  • May 2016 - mycroft-core repositories made public under GPL
  • Jun 2016 - @Aix releases MycroftAI gnome shell extension
  • Jun 2016 - @Aix becomes a contributor on the desktop-integration team
  • Aug 2016 - Steve Penrod becomes MycroftAI CTO
  • 2016

2017

  • Jan 2017 - @Aix release MycrofAI plasmoid for KDE Plasma desktop
  • Mar 2017 - @Aix incubates Mycroft plasmoid project under KDE
  • Apr 2017 - @JarbasAI forks mycroft-core as jarbas-core
  • May 2017 - @JarbasAI becomes a mycroft-core contributor
  • Aug 2017 - MycroftAI starts supporting KDE plasmoid project and installers
  • Sep 2017 - MycroftAI Mark 1 released / mycroft kickstarter campaign fullfilled
  • Sep 2017 - @Aix joins Blue Systems to work on MycroftAI and Plasma integration projects
  • Sep 2017 - jarbas-core tacotron TTS integration added
  • Oct 2017 - mycroft-core relicensed as Apache 2.0
  • Nov 2017 - @Aix becomes a mycroft-core contributor
  • Dec 2017 - jarbas-core abandoned
  • ??? 2017 - Neon AI forks mycroft-core as neon-core to remove dependency on MycroftAI servers and work on other features

2018

  • Jan 2018 - personal-backend repository created, reverse engineered MycroftAI backend by @JarbasAI, licensed as Apache 2.0
  • Jan 2018 - MycroftAI mark 2 kickstarter launch
  • Jul 2018 - personal-backend added to MycroftAI Roadmap
  • Aug 2018 - MycroftAI Mimic2 TTS based on tacotron released
  • Oct 2018 - @j1nx creates "MycroftOS", the first version of what will eventually become the OpenVoiceOS smart speaker
  • Nov 2018 - Mycroft-GUI was created by collaboration between MycroftAI, Blue Systems based on KDE frameworks, maintained by @Aix

2019

  • Jan 2019 - @JarbasAI personal-backend implementation adopted by MycroftAI
  • Mar 2019 - MycroftAI mark 2 completely redesigned and based on different hardware
  • Apr 2019 - Steve Penrod and @JarbasAI create lingua-franca under MycroftAI with @JarbasAI as maintainer
  • Apr 2019 - Chatterbox kickstarter launched by @JarbasAI
  • Jul 2019 - steve penrod leaves MycroftAI
  • Sep 2019 - Mycroft on Plasma Automobile Demo at Akademy 2019 by @aix
  • Oct 2019 - Official MycroftAI backend open sourced, licensed as AGPL
  • Dec 2019 - @Chance joins lingua-franca as a maintainer
  • Dec 2019 - Chatterbox released (closed source) / kickstarter campaign fullfilled
  • Dec 2019 - MycroftAI starts to work on Kivy based UI
  • ??? 2019 - @JarbasAI partners with NeonAI to maintain neon-core

2020

  • Feb 2020 - skill-ovos-common-play forked from Mycroft
  • Mar 2020 - MycroftAI abandons personal-backend
  • Mar 2020 - Michael Lewis becomes MycroftAI CEO
  • May 2020 - @JarbasAI and @AIX partner up to create a VOIP skill
  • Sep 2020 - Community takes stand against reimplementation of GUI using Kivy
  • Oct 2020 - Kivy has been dropped in support for QT Mycroft-GUI
  • Oct 2020 - @JarbasAI, @AIX and @j1nx form OpenVoiceOS community project around mycroft-core
  • Oct 2020 - MycroftOS renamed to OpenVoiceOS - Mycroft Edition due to trademark issues
  • Nov 2020 - @Jarbas gives a talk about translating mycroft for collectivat Jornades de tecnologies lingüístiques lliures en català
  • Dec 2020 - ovos-ww-plugin-snowboy released
  • Dec 2020 - ovos-stt-plugin-vosk released
  • Dec 2020 - ovos-stt-plugin-chromium released
  • 2020
  • ??? 2020 - Chatterbox forks mycroft-core as mycroft-lib (open source)
  • ??? 2020 - mycroft-lib reorganizes mycroft imports cleanly separating each service
  • ??? 2020 - Chatterbox (closed source) recreated on top of mycroft-lib
  • ??? 2020 - neon-core recreated on top of mycroft-lib
  • ??? 2020 - @JarbasAI leaves lingua-franca
  • ??? 2020 - @Chance joins OpenVoiceOS

2021

2022

2023

2024

2025

  • Mar 2025 - ovos-opendata-server introduced: self-hostable, opt-in open-data collection for improving the intent pipeline
  • Mar 2025 - ovos-core reaches the 1.x line (1.1.0), with the modular pipeline architecture stabilized.
  • Apr 2025 - ovos-busmon introduced: web-based live messagebus monitor and debugging tool
  • Jun 2025 - Introduction of ovoscope, the formal end-to-end testing framework for OVOS skills and core components.
  • Jun 2025 - ovos-utils version 0.8.0 released, standardizing cross-platform utilities.
  • Jun 2025 - ovos-workshop walks 4.0.0 → 7.0.0 in one day of breaking cleanups, then drops the skill launcher's compatibility layer for Mycroft-style skill loading two weeks later in 7.0.7a1; ovos-plugin-manager reaches 1.0.0 and ovos-config 2.0.0
  • Jul 2025 - Transcription Validator utterance plugin introduced, filtering STT nonsense before intent matching
  • Oct 2025 - ovos-core reaches the 2.x line (2.1.0), a major release formalizing the multi-intent-pipeline architecture.
  • Oct 2025 - phoonnx introduced: OVOS's own ONNX-based multilingual neural TTS engine and the recommended on-device voice
  • Oct 2025 - OpenVoiceOS selected for an NGI Zero Commons Fund grant by the NLnet Foundation, funding the road to a first stable version, onboarding, language support, and documentation
  • Nov 2025 - Precise wake-word engine ported to ONNX, dropping the tensorflow dependency
  • Dec 2025 - Speech support for minority languages expands: TTS voices for Asturian and Aragonese and the first phonemizer for Barranquenho
  • ??? 2025 - OVOS prototype selected as a demonstrator of the ILENIA project: an open voice assistant in all co-official languages of Spain, runnable on a PC or Raspberry Pi

2026

  • Jan 2026 - ovos-plugin-manager version 2.2.0 released.
  • Jan 2026 - hivemind-core version 4.0.0 released
  • Jan 2026 - ovos-plugin-arena introduced: crowdsourced side-by-side plugin comparison arena
  • Feb 2026 - OpenVoiceOS presents at Speechday 2026, the 4th Dutch Speech Tech Day
  • Feb 2026 - Real-time offline speech recognition lands via ONNX STT plugins
  • Mar 2026 - ovos-installer gains macOS support (Intel and Apple Silicon)
  • Mar 2026 - ovos-agentic-loop introduced: agentic reasoning patterns (ReAct, reflexion, tree-of-thoughts, …) as OVOS agent plugins
  • Mar 2026 - ovos-pydantic-models reaches its first tagged release, packaging the ecosystem-wide data structures first committed in October 2025.
  • Mar 2026 - ovos-bus-client version 1.5.0 released.
  • May 2026 - architecture repo introduced: formal written specifications for the OVOS protocols, with ovos-spec-tools as their dependency-free reference implementation
  • Jun 2026 - MediaProvider plugin family introduced (bandcamp, youtube, tunein, somafm, pyradios, soundcloud, spotify, news, mass) as repositories, intended to eventually supersede the legacy OCP media skills; the in-process loading is not yet wired into ovos-media, see ovos-media
  • Jun 2026 - ovos-test-harness introduced: an executable conformance suite proving the running stack against the formal specs, covering every published spec
  • Jul 2026 - OVOS speech servers gain UTCP/MCP endpoints, callable from any AI agent
  • Jul 2026 - Bus topics migrate to the canonical ovos.* spec namespace, with transparent bridging so legacy and spec components interoperate during the transition — see Bus namespace migration
  • Aug 2026 - The formal specifications start landing in the shipped stack: canonical intent dispatch topics and orchestrator-resident intent context (OVOS-CONTEXT-1) arrive in ovos-core and ovos-workshop

Read next: Reference Overview Related: A Brief History of Open Voice OS · OVOS and Mycroft: A Fork That Wasn't Meant to Be · OVOS Repository Index · Deprecated & Archived Repos