Changelog
Follow up on the latest improvements and updates.
RSS
Our standard avatar model just got a major upgrade.
NEO is the model behind most of the videos you make. We rebuilt it, and the difference is big enough that you will see it in the first video you render. The short version: your avatars now look like a real person even in the moments when they are not talking.
What you will notice
- Pauses look natural.When there is a beat of silence mid-sentence, the face keeps breathing and looking around, the way a person does. It no longer goes blank and waits for the next line.
- The face stays the same face.Across a long take, your avatar keeps looking like the person you picked, instead of slowly drifting into someone slightly different.
- Tricky words hold up.Names, technical terms and fast speech no longer knock the mouth out of shape.
When to use NEO, when to use NEO 2
Same avatar, two models, two different jobs.
- NEOis for when the message is the star. Head and shoulders, calm and precise, the viewer focused on what is being said.
- NEO 2is for when the delivery is the star. Full body, natural gesture and movement, for the moments that need presence and energy.
You are not picking a better or a worse model. You are picking the performance the moment needs.
You don't need to do anything to switch it on. Every video you render with a NEO avatar is already on the new model.
Best,
The Colossyan Team
Three upgrades landed in the editor:
- Scene previews for jump destinations
- Button visual hints for the viewers
- Sharper AI-generated images.
Scene previews for jump destinations.
Setting up a jump target for an Action Button now shows a preview of each scene, instead of clicking through them one by one. Once you're past 30 scenes in a branching video, this saves a lot of hunting around.
A visual cue for where to click.
Action buttons can now show a hint, a small dot that points viewers to exactly where to click. Handy for screen recordings and demo walkthroughs.
Better AI-generated images.
We upgraded the model behind AI image generation in the editor. Expect noticeably better results.
Three upgrades to make the editor smoother and easier to work in.
Best,
The Colossyan Team
Pronunciation control just got a lot better.
You could already fix how words are said in the pronunciation dictionary, but the IPA behind each entry was generated invisibly, with no way to see it or correct it. Now that control is yours: see the IPA, edit it, and preview the result before you commit. It is on by default for everyone, no phonetics background needed.
What changed
- You can now see and edit the IPA.Click a word and a single window opens with every voice variant for it. Adjust the phonetic spelling or the IPA, choose the voice, preview, and save in one place. No more retyping pronunciations into your script.
- AI generation is far more accurate.When you let the editor write the IPA for you, the results land much closer than before, so there is less to fix by ear.
- Voice-specific entries, now spelled out.Each entry applies to the voice you set it for, so switching voices means adding a variant for the new one. This was the most common point of confusion, and the editor now makes it clear.
IPA editing works with US English voices today. More language coverage and polish are on the way.
Open any script, click a word in the pronunciation dictionary, and see the difference.
Thanks for choosing Colossyan, and happy creating!
Best,
The Colossyan Team
You can now choose a cartoon avatar in the editor. It is a fully animated style that sits alongside the realistic avatars, for when you want a character rather than a conventional presenter.
How it works:
- Open the avatars panel in the editor
- Switch on the Cartoon filter to see the animated styles
- Pick a cartoon avatar and add it to your scene
What this means for you:
- Give your training a character and a story, not just a talking head
- Reach for an approachable, animated look when a realistic presenter is not the right fit
- New styles are on the way, so the selection will keep growing
Best,
The Colossyan Team
Finding the right voice used to mean scrolling through hundreds of options. Now you can pin the voices you use most, and they stay right at the top of the list, ready whenever you open the voice selector.
Your most-used voices are always one click away, no matter how large your library gets.
Best,
The Colossyan Team
Your instant avatars just got a serious upgrade.
Instant avatars can now be rendered using NEO 2, our latest avatar rendering model. Until now, NEO 2 was only available for Colossyan avatars. That changes today, bringing the same expressive, lifelike output to the avatars you've built from your own footage.
How it works
When creating an instant avatar, you'll notice a few changes to the wizard:
- Upload a JPG photo to get started. NEO 2 produces the best results from a still image, so we've made that the recommended starting point. The video upload option remains available too.
- The default voice selector has a refreshed layout, though the functionality stays the same.
What this unlocks
- Your custom avatars now match the quality of our best Colossyan avatars
- Branded characters and personas can finally be rendered at NEO 2 quality
Your instant avatars, made better automatically. No reshoot required. 🚀
Best,
The Colossyan Team
Every Colossyan user now works within the same organisation system, previously available to enterprise customers only.
What's changed:
- All users and workspaces now operate under a unified organisation structure
- One organisation owner manages permissions and usage limits across all workspaces in that org
- Workspaces are for structure and asset management. The organisation owner handles access and permissions
A simpler, more consistent experience for every team — fewer edge cases, one place to manage access.
Best,
The Colossyan Team
Eight new avatars are now live and ready to present for you.
Meet our new avatars:
Neha, Kabir, Farhan, Alyan, Sushma, Razza, Zane, and Tanya are now available in the platform
A more diverse roster for teams producing training content for global audiences.
Best,
The Colossyan Team
Your voice library just got more regional. A new batch of high-quality English accents is now available in the editor.
New voices:
- Irish: Mick, Stephen, Shona
- Scottish: Abigail, Chris, Drew, Hector, Skye
What this unlocks:
- Training content that sounds natural and familiar for Irish and Scottish audiences
- Regional authenticity without needing to record your own voiceovers
More regional accents are on the way, shaped by customer feedback and requests.
Best,
The Colossyan Team
Two new content blocks are now available for course creators in Colossyan Learn.
What's new:
- Tables: add structured content directly into lessons — comparisons, step-by-step breakdowns, key data — in a clean, readable format. Rich text editing, header rows, column resizing, and cell merge are all supported
- Diagrams: explain processes, frameworks, and relationships visually inside your lessons
What this unlocks:
- Complex information presented cleanly without switching to external tools
- Learners get a more structured, professional experience inside the course itself
Both are available to all users now.
Best,
The Colossyan Team
Load More
→