With how fast we are pushing out new features, we have to switch back to monthly changelists! The highlight of this month was our recent launch of our self-serve SaaS tier for teams and smaller enterprises, but as always, improvements across the board.
Codeium Teams
Codeium Teams is the most powerful and most secure SaaS coding assistant offering in the market. It is as simple as that. It offers our biggest models, uncapped usage on compute intensive features, in-depth team analytics, and more, all with SOC2 compliance and automatically opted-in zero-data retention guarantees. Read more about our Teams tier for the full list of features, comparisons with other SaaS products like GitHub Copilot, and to see if it is the right plan for you.
More Chat improvements
We continued a lot of our momentum on Chat from August and September into October, and here are some of the highlights.
- Chat in Eclipse and Visual Studio: Following Visual Studio Code and JetBrains, we knew that we needed to bring all of these Chat capabilities to more of our IDEs. Stay tuned for even more of our 40+ Autocomplete-powered IDEs to get Chat as well in the near future.
- Can @ class names: When we launched @ tagging in the previous changelist, it was focused mostly on methods and functions. Now, you can reference to entire classes to pull in relevant context to your Chat queries.
- Multi-conversation history: We introduced multi-window chat in Visual Studio Code in the previous changelist, but made this even cleaner by allowing you to have multiple conversations saved in the same window, so that you can switch between conversations with the AI and preserver the perfect state for each conversation:
Bigger models for everyone
Probably lost in the flurry of visible UX improvements, we have deployed our largest models to date for both Autocomplete and Chat across our SaaS offerings. These models are twice the size of our previous models. Behind the scenes, we did a lot of work on quantization and more infra tricks in order to be able to serve these larger models with the same low latency that you have come to expect from Codeium. So, yes, you are getting better results than you used to, that isn’t your imagination.
Public profiles, usernames, achievement badges, membership cards
We are always motivated when our community helps spread the word about Codeium. Many folks have asked us if there are ways to share how much value they are getting from the tool in order to convince others that they should give Codeium a shot. So, we built out public profiles, where you can select your publicly facing username, and created a way for you to share your high level statistics and achievement badges via a membership card. Just head to your profile page and click the “Share” button on the top right of your analytics to get your membership card link. Thanks in advance for spreading the word!
As always, thank you for being part of the journey. We haven’t mentioned it recently, but if you have not joined our Discord yet, we strongly recommend doing so! Not only is it a great community, but there you will get instructions on how to be the first users of upcoming features, such as GPT-4 support and Codeium Command, a brand new modality separate from Autocomplete and Chat.