Changelog & monthly updates
May 8, 2025
Hi there 👋
It’s Flo here with our monthly update of what we’ve been up to in April (ok, I'm a couple days late, I know 🤷). Here's what's new this month:

💌 New channel: emails
I'm super excited about this guys. Whether, we like it or not, emails are an integrant part of our workflows - for some more than others. I personally rely a lot on emails - I use Superhuman and follow a strict zero inbox policy pretty much daily.
Recently I came to realize how powerful having access to Notis directly within my mailbox could be. Here's the use case that hit me hard and led us to launch this feature today. A few months back, we started hiring for a few different positions (we still are!) and my inbox started getting flooded with candidates' applications. "How cool would it be if I could outsource all this tedious work of sorting out candidates to Notis," I thought.
Now I can and so do you. Here how you can use our new email channel by taking my recruitment example:
Make sure Notis knows your correct email address. You can just ask.Create a new database called "AI engineer applications."Create as many fields as criteria you want to measure candidates against. For each field, add a description to ensure there is no ambiguity, e.g., "True if the candidate has more than 5 years of experience coding in Python," "True if the candidate is within the CET +5 timezone," etc.Create a view to only display the candidates that match your minimum criteria.Start forwarding applications you receive to hey@notis.ai or set a forwarding rule.
For every email you forward, Notis will understand that the email is an application and that you have a database to store those applicants. When trying to create a new entry, Notis will need to complete the fields based on the information selected. Congratulations! You've outsourced candidate filtering to AI.
Of course, this is just an example, and as always with Notis, the possibilities are endless. Other cool stuff that the email channel supports:
When you ask Notis on WhatsApp/Telegram to send you an email, you can now continue the conversation (with all the context) by replying to the email. This is super useful for following up on drafts.You don't need to forward emails for this to work. You can start a normal conversation with Notis by email to do everything you already do with Notis on other channels.Images and documents, whether attached or inline, are fully supported.
As always, I can't wait to hear what you do with this so don't hesitate to shoot me a message (Telegram, Whatsapp) with your favourite use-cases.

🚀 GPT 4.1 and GPT o4-mini
This month, OpenAI launched three new frontier models. We've upgraded all our users to the latest and most powerful models after testing them internally and being very impressed by the increase in intelligence. Here are the models we've selected for you:
GPT 4.1 is replacing GPT 4o as our go-to model for all your agents. It's the most intelligent non-reasoning model to date and also supports vision capabilities.
GPT o4-mini is replacing GPT o3-mini as our planning model. While it's not as intelligent as the new o3, it's also 10 times cheaper, three times faster, and significantly more intelligent in our tests than o3-mini.
This is again a peek into the future of Notis - with or without everything we invest in making our product more useful and more reliable, foundational LLM models will keep getting better, and this will directly impact the level of intelligence demonstrated by Notis.

👾 Our new Discord community and public roadmap
Since we launched, I personally onboarded most of our paid users, and this will continue to be my priority until I feel we fully live up to the capabilities of Notis in terms of video tutorials, documentation, etc. (more on this really soon).
But many of our users asked for a place to gather and exchange ideas, tips, workflows, etc. So this month, we're launching our Discord server. Looking forward to continue the conversation there.
We're also giving you more visibility on our current roadmap and let you vote on the features you want to see us build next.
So there is no better time than now to get your voice heard and help us shape the product you want.
🐞 Other small improvements
You can now pause audio played in our Raycast extension.
Full transcripts are back in Notion documents and emails.
Improved reaction status for Telegram. You can now see exactly which messages are being processed and which have finished processing.
🗞️ In Other News
I recorded an interview with Molly Jones from notionology.com where I share the story behind Notis, my vision and how creators/consulants will play a major role in it. Watch it here!
I published a deep dive comparaison on our blog between Notion AI and Notis and between VoiceNotes.com and Notis
I took a full week off to enjoy my beautiful island. Yay 🎊
Notion lawyers contacted us 😱 to demand that we update our logo, remove the illustrations we own but imitate the style of Notion, remove all Notion logos from the website, and finally ask us to redirect all links to templates to their template store instead of the creator's website. We'll comply with all their requests, even the ones that feel stupid and unfair, because we're just too worried about losing access to their API.
I created an other super cool animation for the website 👏
📰 AI News & Team Picks
Here's everything cool we saw happening this month in AI:
🤯 How, exactly, could AI take over by 2027? Checkout this deeply researched scenario forecast written on intelligence explosion and it's potential consequences for humanity. (X)
🧠 OpenAI Unveils o3, o4-mini and 4.1: OpenAI introduced its most advanced reasoning models to date—o3 and o4-mini and a new non-reasoning model, 4.1. These models enhance capabilities in coding, mathematics, and visual analysis. Notably, o3 can "think with images" and autonomously use tools, marking a significant leap forward in AI reasoning. (X)
🐬 Google Introduces DolphinGemma: Google AI unveiled DolphinGemma, a foundational model trained to understand and generate dolphin-like vocalizations. This initiative aims to decode dolphin communication, bringing us closer to interspecies interaction. (X)
🦙 Meta Launches Llama 4 and AI Assistant App: At its inaugural LlamaCon, Meta announced the release of Llama 4 and a standalone AI assistant app. The app integrates with Facebook and Instagram, offering personalized AI experiences powered by Meta's latest large language model. (X)
🎨 Stability AI Releases Stable Diffusion 3.5: Stability AI introduced Stable Diffusion 3.5, its most powerful text-to-image model yet. The open release includes multiple variants, offering enhanced image generation capabilities for creators and developers. (X)
💰 Elon Musk's xAI Seeks $20B Funding: Elon Musk's AI venture, xAI, is reportedly in talks to raise about $20 billion in funding. The funding round would value the company at over $120 billion, underscoring the growing investor interest in AI-driven enterprises. (X)
That's it for this month. See you in June!
Flo
PS: Do you like what we do? Join our affiliate program and help us spread the word around. You might even make a few bucks in the process 😉