Ekaterina Badaeva
Feb 15, 2026
Third public release
Since the last release, a bunch of useful features have been added and the overall experience has significantly improved. Our team is excited for you to hear about them, and we hope you are too!
Let’s start with the Major Changes:
Very often, when working with AI models, we don’t want a single answer - we want to compare. How do different models answer the same question? Which direction feels more promising?
So, instead of forcing you to pick one model upfront, we redesigned the experience around comparison. You can now send requests to multiple models at the same time.
Multiple Models
Very often, when working with AI models, we don’t want a single answer - we want to compare. How do different models answer the same question? Which direction feels more promising?
So, instead of forcing you to pick one model upfront, we redesigned the experience around comparison. You can now send requests to multiple models at the same time.
Prompt node sending requests to multiple models, branching into a tree
The flow visually branches in different directions, forming a tree. From there, you can either continue exploring with all models or branch off by clicking the arrow button on a content node to develop a specific path further.
You can quickly create a prompt node with a default model preset (XS / S / M / L), so you can start experimenting instantly without overthinking configuration, while still having the option to manually select up to 4 models.
We also expanded file support - you can now import files via drag & drop, including PDFs and audio. But not just import them - actually use them:
You can quickly create a prompt node with a default model preset (XS / S / M / L), so you can start experimenting instantly without overthinking configuration, while still having the option to manually select up to 4 models.
File Support
We also expanded file support - you can now import files via drag & drop, including PDFs and audio. But not just import them - actually use them:
- Transcribe audio with Whisper.
- Extract text from PDFs and images using the Vision capabilities of models.
Importing PDF and audio files and converting them to text
Space for canvases
Everyone knows that one flow is rarely enough - different ideas, contexts, or experiments deserve their own space to stay organized and clear.
So now you can create and manage multiple canvases inside a dedicated Space. Switch between grid and list layouts, duplicate canvases, rename them, export them, or remove the ones you no longer need.
Multiple canvases in Space view with grid layout
Templates
And to make onboarding and experimenting smoother, we introduce Templates.
Explore ready-to-use flows prepared by our team, add them to your canvas, or export them for your own use.
Explore ready-to-use flows prepared by our team, add them to your canvas, or export them for your own use.
Library of templates available to add to the canvas
Now, about smaller improvements:
- No more manual saving! Your flow is now saved automatically.
- New nodes get auto-focused right after requests or when you insert a template.
- You can now upload your avatar to personalize your experience.
- Contact info was added in the About section so it’s easier to stay in touch with us.
Besides these updates, there are many minor UI and logic improvements that make everything feel faster, cleaner, and more consistent.
We continue simplifying complex workflows and shaping Devneya into something powerful and still intuitive. We’re excited to hear what you think!
Sincerely yours,
-- Devneya team.
We continue simplifying complex workflows and shaping Devneya into something powerful and still intuitive. We’re excited to hear what you think!
Sincerely yours,
-- Devneya team.