> For the complete documentation index, see [llms.txt](https://docs.atlas.design/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.atlas.design/atlas-ai-studio-overview/release-notes/release-0.52.0.md).

# Release 0.52.0

### 1. OpenAI GPT-Image-2.6 Sunburst and Flare&#x20;

Atlas now supports OpenAI's flagship **GPT-Image-2.6 Sunburst** and **GPT-Image-2.6 Flare** models across **Text to image, Multimodal**, and **Multimodal Array**.

* **Model Tiers:**
  * **GPT-Image-2.6 Sunburst:** Flagship high-detail image generation and multi-reference editing with crisp typography, lighting, and material fidelity.
  * **GPT-Image-2.6 Flare:** Fast, cost-efficient image generation and reference-guided editing for rapid concept iteration.
* **Resolutions & Aspect Ratios:** Supports **1K**, **2K**, and **4K** output resolutions across `1:1`, `3:2`, `2:3`, `16:9`, and `9:16` aspect ratios, plus `match_input_image` and `rough_match_input_image` when reference images are connected in **Multimodal** and **Multimodal Array**.
* **Reference Images:** Accepts up to **10 reference images** in **Multimodal** and **Multimodal Array** for character, style, and composition consistency.
* **Fine-Grained Controls:** Configure **Quality** (`low`, `medium`, `high`, `xhigh`, `max`) and **Complexity** (`low`, `medium`, `high`) to balance visual detail, reasoning depth, and credit usage.

<figure><img src="https://3654894688-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FR7boiMixMhR4q36Ns33Y%2Fuploads%2FLVUGgYxccPg4MNJ1XPtD%2Fimage.png?alt=media&amp;token=2e197611-b8df-4649-b101-567b03f056cf" alt="" width="433"><figcaption></figcaption></figure>

<figure><img src="https://3654894688-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FR7boiMixMhR4q36Ns33Y%2Fuploads%2FpX3hdGAQwb7KKOcc5ghv%2Fimage.png?alt=media&amp;token=240db80b-c842-4c5e-8bac-a8c6db0b259d" alt=""><figcaption></figcaption></figure>

***

### 2. ByteDance Seedream 5.5&#x20;

**ByteDance :: Seedream 5.5** (powered by native BytePlus) is now available in **Text to image**, **Multimodal**, and **Multimodal Array**.

![ByteDance Seedream 5.5 Text to Image and Multimodal Workflow](https://3654894688-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FR7boiMixMhR4q36Ns33Y%2Fuploads%2FRKQiqfR9krg8Btp3vSe5%2F14-seedream-5-5-workflow.jpg?alt=media)

***

### 3. Anthropic Claude Fable 5.2 (`Text generation (LLM)`, `Image to text`, `Structured output (LLM)`, `Bounded Agent`)

**Anthropic Claude Fable 5.2** is now the default flagship Claude model across Text generation (LLM), Image to text , Structured output (LLM), and Bounded Agent.

* Adaptive Reasoning Effort: Choose Reasoning (`None`, `Low`, `Medium`, `High`, `xHigh`, or `Max`) to control thinking depth across creative writing, visual scene analysis, structured JSON extraction, and multi-turn agentic tool use.

![Anthropic Claude Fable 5.2 LLM and Vision Workflow](https://3654894688-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FR7boiMixMhR4q36Ns33Y%2Fuploads%2Fc3JYvvguzz4c6cwv8t74%2F15-claude-fable-5-2-workflow.jpg?alt=media)

***

### 4. Pruna P-Video-2 Pro Replace Person Swap in `Edit video`

The **Edit video** node now includes **Pruna :: P-Video-2 Pro Replace** for high-consistency character and person replacement in video clips.

* **1–2 Person Replacement:** Supply a source video (`MP4`, `MOV`, or `WebM`) and **reference photos** via **New People Photos \[P-Video Replace]** to swap one or two subjects while preserving original camera motion, scene lighting, and body performance.
* **Output Resolutions up to 4K:** Select **Replace Resolution** (`480p`, `768p`, `1080p`, `2k`, or `4k`).
* **Lip-Sync & Prompt Guidance:** Toggle **Ignore Source Audio For Lip Sync**  when replacing silent or non-speaking subjects.

![Pruna P-Video-2 Pro Replace Person Swap Workflow](https://3654894688-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FR7boiMixMhR4q36Ns33Y%2Fuploads%2FVoYroZJumS100QEfJLPq%2F03-p-video-2-pro-replace-workflow.jpg?alt=media)

{% embed url="<https://files.gitbook.com/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FR7boiMixMhR4q36Ns33Y%2Fuploads%2F4bEJdzkYOALDC7AesP3g%2Fp-video-replace-source-5a208b6f.mp4?alt=media>" %}

{% embed url="<https://files.gitbook.com/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FR7boiMixMhR4q36Ns33Y%2Fuploads%2FTF9HE8FOYJ64jGTJfRFs%2Fp-video-replace-cb5cbd5c.mp4?alt=media>" %}

***

### 5. Pruna P-Video-2 Pro Reference Generation in `Text + image to video`

The **Text + image to video** node now includes **Pruna :: P-Video-2 Pro Reference**  for multi-reference character- and prop-consistent video generation.

* **Reference Images + Optional Keyframes:** Connect **Reference Images** to anchor character identity, props, or environment style, along with an optional **First Frame**and **Last Frame** .
* **768p to 4K Resolutions:** Generate **1–10 second** clips at **768p**, **1080p**, **2K**, or **4K** across `16:9`, `9:16`, `4:3`, `3:4`, and `1:1` aspect ratios.
* **Mode & Optimization Controls:** Choose **Mode** (`speed` or `cost`), **Prompt Expansion** (`off`, `turbo`, or `max`).

![Pruna P-Video-2 Pro Reference Workflow in Text + Image to Video](https://3654894688-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FR7boiMixMhR4q36Ns33Y%2Fuploads%2FYKU4ew6hJTXTIGU5ppwP%2F16-p-video-2-pro-reference-workflow.jpg?alt=media)

{% embed url="<https://files.gitbook.com/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FR7boiMixMhR4q36Ns33Y%2Fuploads%2Fv6PZMt2J2e3tcy7BaN50%2Fp-video-2-pro-reference-79f2b003.mp4?alt=media>" %}

***

### 6. Alibaba Wan 3.0 Edit Beside Live Wan 2.7 Edit in `Edit video`

**Edit video** now offers **Alibaba :: Wan 3.0 Edit** alongside the existing **Alibaba :: Wan 2.7 Edit** engine for reference-conditioned video restyling, character edits, and audio-synchronized transformations.

* **480p & 720p & 1080p Output Resolutions**
* **Character Reference Images:** Connect to anchor appearance, wardrobe, or scene aesthetics across the edited clip.
* **Audio, Prompt Rewriting & Thinking Controls:** Toggle **Audio \[Wan 3.0]**, **Rewrite Prompt \[Wan 3.0]**, and **Think First \[Wan 3.0]** under Advanced controls.

<figure><img src="https://3654894688-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FR7boiMixMhR4q36Ns33Y%2Fuploads%2FE2txQ5wdqZHvIJ4YsyPq%2Fimage.png?alt=media&amp;token=e66520f5-0936-40ec-a756-0396b575c4f5" alt=""><figcaption></figcaption></figure>

{% embed url="<https://files.gitbook.com/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FR7boiMixMhR4q36Ns33Y%2Fuploads%2FAG2Fu6sCDJPs3yLY6ClO%2Fedited_video%20(10).mp4?alt=media&token=30b65bc9-9308-4a18-bc81-f458e7b66295>" %}

***

### 7. Vidu Q4 Preview Video Generation in `Text + image to video`

The **Text + image to video** node now supports **Vidu :: Q4 (Preview)** alongside **Vidu :: Q3**, offering both first-frame animation and multi-reference video generation with separate video and cover image outputs.

* **Two Generation Modes (`Q4 Mode`):**
  * **First Frame:** Animate a connected **First Frame.**
  * **References:** Combine **reference images** and **MP3 voice clips** across `16:9`, `9:16`, `4:3`, `3:4`, and `1:1` aspect ratios with optional audio generation.
* **Dual Output Ports:** Emits both **Output Video** and a dedicated **Cover Image** port for thumbnails and downstream frame chaining.

![Vidu Q4 Preview First Frame Workflow](https://3654894688-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FR7boiMixMhR4q36Ns33Y%2Fuploads%2F30FyuPiXQTckDPB1aFmv%2F13-vidu-q4-preview-workflow.jpg?alt=media)

<figure><img src="https://3654894688-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FR7boiMixMhR4q36Ns33Y%2Fuploads%2F2iCJB5R4182ZreicJ7zj%2Fimage.png?alt=media&amp;token=4245887c-a9c6-4ea9-a40d-a28a009e501a" alt=""><figcaption><p>Vidu Q4 Preview References Workflow</p></figcaption></figure>

***

### 8. LTX-2.5 Pro & Fast Audio-to-Video (`Driving Audio` in `Text + image to video`)

The **Text + image to video (Simple)** node now accepts an optional **Driving Audio**  input when **LTX-2.5 Pro**  or **LTX-2.5 Fast** is selected, turning any voice track, music stem, or sound effect into a synchronized **1080p** video clip.

* **Audio-Driven Timing & Framing:** Connect an audio clip on **LTX-2.5 Pro** or **LTX-2.5 Fast** (with an optional **Reference / First Frame** image). Output duration automatically follows the driving audio.
* **Streamlined Controls:** Connecting **Driving Audio** automatically hides non-applicable fixed-duration controls (`Duration`, `Resolution`, `Frame Rate`, `Camera Motion`, `Disable Audio`, and `End Image`) so only the active audio-to-video parameters are shown.

<figure><img src="https://3654894688-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FR7boiMixMhR4q36Ns33Y%2Fuploads%2Fn2pBW136iyBVPWA5CVp8%2Fimage.png?alt=media&amp;token=377e2b2c-c51c-4781-a231-d56e8049d061" alt=""><figcaption></figcaption></figure>

<figure><img src="https://3654894688-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FR7boiMixMhR4q36Ns33Y%2Fuploads%2FH3Eowgc8X3XSKWyxIxiT%2Fimage.png?alt=media&amp;token=5e84f48d-b23d-48ef-8181-7fc6ffc57768" alt=""><figcaption></figcaption></figure>

{% file src="/files/snHoFIUWG6loKnHCvweK" %}

{% embed url="<https://files.gitbook.com/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FR7boiMixMhR4q36Ns33Y%2Fuploads%2FArCO8kI28vKf4lW8TP9U%2Fltx25_audio_video.mp4?alt=media&token=4405ff4b-2a4c-4e3e-985b-cefa5e24bb41>" %}

***

### 9. MiniMax Manga Animation, 360° Orbit & Card Spin Node (`MiniMax Manga Animation`)

A new **MiniMax Manga Animation** node is now available under **Video → Generate** to animate 2D illustrations, manga panels, and character sheets in three specialized LoRA-driven modes, and **Video camera orbit** now shares the unified MiniMax H3 LoRA wrapper and fal receipt billing:

* **Three Animation Modes (`mode`):**
  * **Manga Action Animation (`manga_animation`):** Brings linework and manga panels to life while preserving ink and halftone styling.
  * **Manga 360° Orbit (`manga_orbit`):** Performs a full 360-degree camera turnaround around the illustrated subject using rescaled `ORB360` camera milestones and locks the source drawing as the end frame.
  * **Manga Card Spin (`card_spin`):** Spins the illustration edge-on like a collectible card with `ORB360` card-spin triggers and end-frame locking.

<figure><img src="https://3654894688-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FR7boiMixMhR4q36Ns33Y%2Fuploads%2FzoaaXFZczxgl54RliUUR%2Fimage.png?alt=media&amp;token=b42e5fb9-294d-4579-bd75-89fd1f587ed8" alt=""><figcaption></figcaption></figure>

* **Flexible Duration, Resolution & LoRA Strength:** Supports **5–15 second** clips at **480P**, **720P**, **1080P**, **2K**, or **4K** resolution with adjustable **LoRA Scale** (`0–4.0`) and **Prompt Expansion** (`fast`, `balanced`, `quality`).

<figure><img src="https://3654894688-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FR7boiMixMhR4q36Ns33Y%2Fuploads%2FFpMceHXkFmkqjwHUMvu6%2Fimage.png?alt=media&amp;token=520214e7-3f06-4efa-bf97-382b02195800" alt=""><figcaption></figcaption></figure>

{% embed url="<https://files.gitbook.com/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FR7boiMixMhR4q36Ns33Y%2Fuploads%2FPMXnf66eRsS2J1IcEDY0%2Fmanga_animation.mp4?alt=media&token=6fbec4f8-a8f8-46ee-9844-a5013fea036a>" %}

***

### 10. Seedance 2.5 Pro Camera-Guided Video Node (`Camera-guided video`)

A new **Camera-guided video** node is now available under **Video → Generate** (right after **Reference to video**) to generate shots whose visual look comes from a **Start Frame** and whose camera trajectory follows a **Camera Previs Video** (such as a Blender viewport playblast or blocking pass).

* **Automatic Reference Role Binding:** Connects **Start Frame** as `@image1` and **Camera Previs Video** as `@video1` (used strictly for camera motion and spatial blocking), so your **Action** prompt only needs to describe what happens in the scene without manual `@image1`/`@video1` tokens.
* **4–30 Second Shots up to 1080p:** Configure **Duration** (`4–30s`, independent of previs clip length), **Resolution** (`480p`, `720p`, `1080p`), **Aspect Ratio** (`16:9`, `9:16`, `4:3`, `3:4`, `21:9`, `1:1`), **Generate Audio** (on by default), and **Non-celeb faces in ref**.

<figure><img src="https://3654894688-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FR7boiMixMhR4q36Ns33Y%2Fuploads%2FgIXazaciZ5xahyIULqir%2Fimage.png?alt=media&amp;token=c78036d2-8e0c-4a3b-bcec-fe9fcf4f0e3f" alt=""><figcaption></figcaption></figure>

{% embed url="<https://files.gitbook.com/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FR7boiMixMhR4q36Ns33Y%2Fuploads%2Fsb3KZziXXbrWZTuM8vyN%2Fcamera_orbit.mp4?alt=media&token=b064f1a9-5293-4656-9be2-43c9c182379b>" %}

{% embed url="<https://files.gitbook.com/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FR7boiMixMhR4q36Ns33Y%2Fuploads%2FaUtAc9a12BOv5tFcjeSI%2Fcamera_guided_video.mp4?alt=media&token=fc7298bb-5424-4d4e-a3f2-1b7d1826cec0>" %}

***

### 11. Frame to Motion: One-to-All Animation Node (`Frame to Motion`)

A new **Frame to Motion** node is now available under **Video → Edit** (powered by `fal-ai/one-to-all-animation/14b`) to transfer full-body human or character motion from a **Driving Video** onto any **Reference Image**.

* **Pose & Motion Transfer:** Connect a **Reference Image** and a **Driving Video** with a descriptive **Prompt** to animate characters, illustrations, or concept art following the driving performance.
* **Resolution, Frame Rate & Alignment Controls:** Choose **Resolution** (`480p`, `580p`, or `720p`), **Identity Guidance Scale** (`0.0–10.0`), **Motion Guidance Scale** (`0.0–10.0`), **Negative Prompt.**

![Frame to Motion One-to-All Animation Workflow](https://3654894688-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FR7boiMixMhR4q36Ns33Y%2Fuploads%2FtBfEUdiLcLH5hxJra0SL%2F18-frame-to-motion-workflow.jpg?alt=media)

{% embed url="<https://files.gitbook.com/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FR7boiMixMhR4q36Ns33Y%2Fuploads%2FlUnqYyAIBCLIUoQXl2PH%2Fframe-to-motion-206209da.mp4?alt=media>" %}

***

### 12. Meshy Image to Relief Node (`Image to relief`)

A new **Image to relief** node is now available under **3D → Generate** to convert a 2D illustration, emblem, coin, or architectural motif into both a 3D bas-relief mesh and a 16-bit depth map.

* **Dual 3D Mesh + Depth Map Outputs:** Emits both a watertight 3D **Output Mesh** and a **16-bit PNG Depth Map** for displacement shaders, normal map baking, or CNC/3D-printing workflows.
* **Sculpting & Geometry Controls:** Configure Carving Mode, **Thickness** (`0.01–50.0`), **Enhance Details**, and **Remove Base**.

<figure><img src="https://3654894688-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FR7boiMixMhR4q36Ns33Y%2Fuploads%2FDo0i8xI35GDRYZvJ47Do%2Fimage.png?alt=media&amp;token=ada15893-4b3c-452a-bbcc-ab836202b4e6" alt=""><figcaption></figcaption></figure>

![Meshy Image to Relief 3D Bas-Relief and Depth Map Workflow](https://3654894688-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FR7boiMixMhR4q36Ns33Y%2Fuploads%2F0l130knk6SBrEwXfkOtS%2F17-meshy-relief-workflow.jpg?alt=media)

***

### 13. ElevenLabs SFX v3 with Seamless Looping in `Sound effects`

The **Sound effects** node now defaults to **ElevenLabs :: SFX v3** alongside **ElevenLabs :: SFX v2**.

* **Seamless Loop Toggle (`loop`):** Enable **Loop** when **ElevenLabs :: SFX v3** is selected to generate seamlessly looping ambient beds, engine hums, and environmental soundscapes across **0.5–22 second** durations.

![ElevenLabs SFX v3 Sound Effects Workflow](https://3654894688-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FR7boiMixMhR4q36Ns33Y%2Fuploads%2FMDJkCJv51ANTAI6zRADt%2F19-elevenlabs-sfx-v3-workflow.jpg?alt=media)

***

### 14. Interactive 3D Mesh Turntable Previews on Canvas & Asset Cards

3D mesh nodes on the canvas (`Input 3D`, `Image to 3D`, `Image to relief`, `(Re)Texture mesh`, etc.) and 3D assets in the asset picker now render live, interactive **3D turntable previews** right on the card instead of flat static PNG thumbnails.

* **Pointer-Push Rotation & Draggable Degree Scale:** Move the pointer across a 3D card or drag the bottom degree scale (`R · 315 · F · 45 · L`) to spin the mesh with momentum, accompanied by a live cardinal orientation readout (`Front · +Z`, `Left · +X`, `Back · -Z`, `Right · -X`, and exact degrees) and a `"Click the preview to open 3D. Drag the scale to rotate."` tooltip. Clicking the preview opens the full **3D Viewer** modal.
* **User Preferences (`Profile → Hotkeys & UI`):** Toggle **Animate 3D previews** (`animateMeshPreviews`) to switch between interactive 3D turntables and static thumbnails, or toggle **Invert 3D preview scroll** (`invertMeshPreviewScroll`) to reverse drag/push rotation direction.

{% embed url="<https://files.gitbook.com/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FR7boiMixMhR4q36Ns33Y%2Fuploads%2Fq0YIt5hpMPxeIT5VogdW%2F20261112-0947-01.3585128.mp4?alt=media&token=4e08bed4-bdf5-4e94-b968-46f1e47f6103>" %}

***

### 15. Native GIF Format Support (`Input GIF`, `Disassemble GIF`, `Assemble GIF`, and Exported APIs)

Animated GIFs are now a first-class media type (`gif`) across the Atlas canvas, asset picker, and exported APIs, complete with dedicated cyan ports and wires, native looping canvas previews, and frame-level disassembly and reassembly nodes.

* **Input GIF (`input_gif`):** Upload or select animated `.gif` assets from your workspace catalog and preview them looping natively on the canvas without static thumbnail conversion.
* **Disassemble GIF (`gif_to_images`):** Extract up to **300 full-frame composited RGBA PNG frames** in playback order while preserving transparency, along with the detected average playback speed (**FPS**, bounded to `1–50`) ready to wire directly into **Assemble GIF**.
* **Assemble GIF (`images_to_gif`):** Combine up to **300 same-sized image frames** into an animated GIF with configurable **FPS** (`1–50`) and a **Loop** switch (`Loop = true` repeats forever; `Loop = false` plays once).
* **Exported API Support (`api_input_gif` & `api_output_gif`):** Build and export automated GIF processing pipelines via **API Input GIF** and **API Output GIF** and execute them through `/0.2/upload`, `/0.2/api_execute`, and `/0.2/download_binary_result`.

{% embed url="<https://files.gitbook.com/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FR7boiMixMhR4q36Ns33Y%2Fuploads%2FmNU9yo9yMgyaRbbeNVPz%2F20261112-0949-22.3147349.mp4?alt=media&token=68725f51-4362-414c-bdd9-f5ab85f1d630>" %}

***

### 16. Unified Explore Page (`Showcases`, `Templates`, `New Models`, `Atlas Labs`, and `Tools`)

The sidebar navigation now consolidates **Templates** and **Labs** into a single **Explore** destination (`Projects`, `Explore`, `Assets`, `API`), bringing every way to start or extend a workflow onto one searchable page:

* **Showcases:** Browse curated production results—click **Recook** to jump straight into the Lab that created an asset, or **See how** to inspect the underlying node graph read-only.
* **Templates & New Models:** Launch ready-made node workflows or test newly added model templates with a single click (`Open` / `Open your copy`).
* **Atlas Labs:** Access **Quick Labs**, **Lab workspaces**, and **Build a Lab** directly from Explore.
* **Use Atlas in your tools:** Connect **Claude Code** (MCP), **Unity** and **Unreal** engine plugins, or the **Atlas API** from the bottom integration shelf.
* **Instant Search:** Filter across templates, Labs, and models simultaneously from the top search bar.

<figure><img src="https://3654894688-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FR7boiMixMhR4q36Ns33Y%2Fuploads%2Fthv5VpvPd8XmlQC9lob7%2Fimage.png?alt=media&amp;token=b74b74af-62c3-4d0b-9a96-12385227bd88" alt=""><figcaption></figcaption></figure>

***

## Fixes & Polish

### 17. 3D Viewer Display Modes, Persistent `Surface grid` Toggle, Double-Sided Beauty Render, UE5 Rig Framing, and Blender 5.2.3 Pipeline Compatibility

* Display modes: The 3D viewer now offers Beauty, Albedo, Geometry Normals, and Just Wireframe (plus Part Colors for multi-part meshes). A new Surface grid switch lets you overlay the wireframe on any mode, with its own color picker, and the viewer remembers your choice between sessions.
* Cleaner Beauty render: Meshes now render correctly from every side, so open shells, foliage, and thin geometry no longer vanish or flicker when you rotate the model.
* Better character framing: Unreal Engine 5 Mannequin characters are now centered and zoomed correctly when they load and when you click Fit to View.

<figure><img src="https://3654894688-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FR7boiMixMhR4q36Ns33Y%2Fuploads%2FdnUkyavixeaJakXwqitO%2Fimage.png?alt=media&amp;token=ceabf66d-8d5e-4304-8c32-c1f5d377eab5" alt=""><figcaption></figcaption></figure>

***


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs.atlas.design/atlas-ai-studio-overview/release-notes/release-0.52.0.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.
