Updates

Dear Creator,

Canvas is officially live! 🎉 A brand-new creative experience awaits:

• Premium AI Models — Seedance 2.0, Kling, NB Series, MJ, Vidu, Pixverse and more top closed-source models, with more coming soon
• All-in-One AI Design — A professional suite with more models and tools, from idea to final product seamlessly
• Infinite Canvas — An endless creative space with no boundaries

---

💡 Important Notes

• New Resource System: Canvas runs on an independent resource called Energy, completely separate from your TA Credits. Energy is used exclusively for Canvas generation.
• Model Migration: Select closed-source models will gradually move from TA to Canvas exclusives. Seedance 2.0 is Canvas-exclusive starting today. NB Series and Wan2.7 will follow within two weeks. Plan accordingly.

---

🎁 Double Launch Rewards (Apr 13 – 27)

1. Invite & Earn — Gather friends, earn massive free Energy
2. Credits → Energy — Convert your TA Credits to Energy and start creating instantly

❗ Energy pool is limited — first come, first served!

👉 Event details: https://tensor.art/blackboard/canvas-go

👉 Start creating: https://tensor.art/canvas

🆕 Canvas

Model Updates: HappyHorse is coming. ✨ Text → Video & Image → Video 🎯 Highly Responsive to Prompts: You Get Exactly What You Ask For
🎬 Motion is controllable; footage is ready for immediate use
🔗 Consistent across frames; the narrative holds together try now👉https://tensor.art/canvas

🆕 Canvas

Model Updates: Wan2.7 added & Kling gets motion control.

Feature Drop: Support for zooming into single tasks.

try now👉https://tensor.art/canvas

🔧 Service Notice

The AD capabilities of Z-Image-Turbo, Z-Image, and Anima are currently not performing as expected and are under maintenance.

We have resolved the AD issue with the Illustrious model. However, due to the unique characteristics of their training data, we need to re-adjust the AD architecture for the ZIB, ZIT, and Anima models.

Our team is actively working on restoring full functionality. We appreciate your patience and understanding during this time.

Stay tuned for updates on service restoration.

🚀 Exciting News!

Vidu Q3 Pro and PixVerse v6 are launching on Canvas today!

Get ready to create stunning videos with next-gen quality and faster performance. Both models are now available for your projects.

Start creating now and explore what these powerful tools can do.

We’ve optimized how posts are viewed — clicking a post now opens it in a pop-up window instead of navigating to a new page.

Why? Previously, going back from a post detail page could cause unexpected refreshes or lose your scroll position in the feed. The pop-up keeps you right where you were.

Tip: If the pop-up feels too small, just copy the URL and open it in a new tab for the full-screen experience.

Core Highlights

• 🧑‍🎨 Unique Faces: Fine-grained character sculpting — face shape, bone structure, eyes — finally, AI portraits that look like real people.
• ✍️ Print-Level Text: Up to 3K tokens, 12 languages, tables & formulas rendered with print clarity.
• 🖼️ Full Workflow: Text-to-image / instruction editing. 1K–4K + native transparent PNG.
• 🤝 Multi-Subject Consistency: Up to 9 reference images — group shots, posters, composites stay on-model.

Try it now:

https://tensor.art/models/982520730400887123

Incident Notice

We are currently experiencing a service disruption due to issues
with Google Cloud Platform (GCP). As a result, Nano Banana
related features may be temporarily unavailable or degraded.

Affected Features

• Nano Banana image generation

What You Can Do

Please try again later. Other features on the platform remain fully operational.

We apologize for any inconvenience and will update this notice once the issue is resolved.

Homepage Revamp

The homepage now features 10 purpose-built AI creation tools, each designed for a specific creative
workflow:

• OC Creator — Design original characters from scratch
• Anime Lab — Anime characters, scenes, and stylized animations
• Photoreal Studio — Generate photorealistic images
• Seamless Video — Start from any frame for continuous, uninterrupted video
• Smart Edit — One-click photo retouching, transformation, and enhancement
• Create from Reference — Generate images/videos based on reference photos
• Digital Avatar — Create your own virtual digital doppelgänger
• Motion & Pose Ref — Full control over poses, camera angles, and movements
• SNS Trends — Instantly recreate viral social media trends
• Brand Design — Your one-stop solution for logos and visual branding assets

No need to master complex parameters — simply pick a tool and start creating!

What Changed

We identified an issue in the new post publish UI where the image reordering controls became unavailable when a post contained more than 4 images. This made it impossible to adjust the display order before publishing.

We have made a UI optimization to resolve this — image order can now be adjusted regardless of how many images are included in the post.

Notice

This is because video tasks only utilize the accelerated generation pipeline when the CFG value is set to 1; if the CFG were to be adjusted to 2 while in Fast Mode, the generation time would increase significantly—thereby defeating the very purpose of Fast Mode.

Apology

We sincerely apologize for the failure to provide timely notification of this change. A breakdown in our internal information synchronization mechanism resulted in this update not being released to the public immediately.

What Happened

Civitai has discontinued their model transfer stream, which our import feature relied on. As a result, importing models directly from Civitai is temporarily unavailable.

What You Can Do

• Upload locally — Download your model files from Civitai and upload them directly through the local file upload option

What’s New

We’ve added hash-based model search to Tensor Art. Simply type @sha256 in the
search box followed by the hash value to quickly locate the exact model you’re looking for.