At long last, I am very proud to announce that Almond 2.0 is now available. This is a culmination of several months of development since 1.99, and it’s a major update compared to the previous stable version of Almond 1.8.
Compared to Almond 1.8, this release brings:
- better support for the most important skills: Home Assistant, music, weather
- more robust natural language, including a new, state-of-the-art conversational model that can understand follow-ups and refinement of queries
- for developers, the release includes a newly repackaged Genie Toolkit, with ThingTalk 2.0. The new toolkit is easier to learn and use, and more robust.
As usual, you can read more about this release in the release page in the wiki .
Many of you have seen early previews and development releases leading up to this point. Compared to the beta, we have refreshed our UI and fixed some of the issues that you reported. Thank you all for helping the Almond project up to this point!
Not all skills that were part of Almond 1.8 are available, as we focus on a smaller set of higher-quality skills. The following skills are fully supported:
- Bing Search
- Yelp restaurants
- Home Assistant
Additional skills are supported by the community, and will be released over time. These currently include Xkcd, quotes, cat and dog pictures, RSS.
The release will go live on the https://almond.stanford.edu server tomorrow (Thu May 27) at 9am Pacific. As it is a major release that involves substantial backend changes, please expect potentially several hours of downtime.
I will post an update here when the deployment is complete, or if there issues that can prolong downtime.
After the new version of the backend servers is available, we will release the builds of the Almond client applications:
- for Home Assistant, on the official add-on repository (superseding the edge version on the unstable repository)
- for Flatpak/GNOME, on Flathub
If you encounter any issue with the release, please let us know. You can reach us here in he community forum, on Discord, or by filing an issue on GitHub.
If you file an issue, please include the detailed conversation log.