DocsMedia Studio
Images and 3D models in Media Studio
Media Studio makes images and 3D models from a prompt or reference pictures, with your own provider keys. Results wait in its library until you add them to your scene.
Opening Media Studio#
- Studio, beside Chat in the top bar, in Director and Manual. Click it again to close Media Studio, or in Manual to go back to the workspace you came from.
- + at the left of the message box ▸ Media Studio, or Video ▸ Video for the Video tab, or Workflows.
- Media Studio at the top of the Angelo menu, and the Media Studio pill in the Asset Browser and on asset shelves.
- In Manual, the Media Studio workspace tab, or any area's Editor Type menu ▸ AI ▸ Media Studio.
In Director, Media Studio opens in its own window, so your layout is never rearranged.
Connect a provider#
Media Studio uses its own API keys, separate from the chat's accounts. Add them in Preferences ▸ AI Providers ▸ Media generation (API keys). If you are signed in to a team, its keys appear under the team's name.
| Connection | Makes | Models |
|---|---|---|
| OpenAI Images | Images | GPT Image 2.5 Flare, GPT Image 2.5 Sunburst |
| Gemini Images | Images | Nano Banana 2.1, Nano Banana 2, Nano Banana Pro |
| BytePlus ModelArk | Images and video | Seedream 5.0 Pro; Seedance video (see Video generation) |
| Tripo 3D | 3D models | Tripo P2, Tripo 3.1 |
| Meshy 3D | 3D models | Meshy 7, Meshy 7.1 |
| fal | Video and 3D models | Video models (see Video generation); TRELLIS 2 |
Nano Banana 2.1 appears when your Gemini key offers it; otherwise Nano Banana 2 is the default.
Make an image or a 3D model#
- Choose the Image or 3D tab.
- Pick the model. The picker names each model's provider and the account it bills. Under it, the Create column names that account, or says which one is missing, with Connect in AI Providers.
- Write a prompt. Expand opens a bigger editor, and AI Enhance rewrites your prompt: you review the result before it replaces yours.
- Add references if you want them: Image for a file, URL for a public web address, or drop pictures on the editor. An image takes up to four. A 3D model takes a prompt or one picture, except TRELLIS 2 (below).
- Set the model's options: aspect ratio and resolution for images, a target polygon count and textures for 3D. Only the options that model supports are offered.
- Click Generate image or Generate 3D. A confirmation names the model and account, who bills you, and how many pictures will be uploaded; Meshy's textured models from a prompt say they bill two stages. It runs when you click Generate.
Up to four generations run at once, and nothing is retried automatically. Stopping a generation only stops Angelo watching it: a task the provider already received may still finish and be billed.
You can also ask in the chat: the agent can generate an image or a model for you, and it asks first. A model it makes can go into your scene; an image goes to Media Studio's library, like one you make yourself. To chain steps, for example an image that becomes a 3D model, use Generation Flows; Turn into a flow, at the right of Media Studio's top row, starts one from your current setup.
TRELLIS 2 through fal#
New in 1.0.8TRELLIS 2 makes a textured 3D model from pictures. It is in the 3D tab beside Tripo and Meshy once a fal key is connected (AI Providers ▸ Media generation ▸ fal).
- It works from pictures only: one image, or up to four views of the same object. A prompt alone isn't enough.
- It always textures its model, so Textures stays on.
- The target polygon count can be from 5,000 to 2,000,000.
TRELLIS 2 is offered wherever a 3D model is: in Generation Flows, when you drop a photo on the 3D view (Photo to 3D model), and to the agent.
Prices and the paying account#
Every generation is billed by the provider to the API account the model belongs to. Removing an account never moves its models to another account. A team's key says so in the confirmation: it uses your team's provider key through Angelo.
Where Angelo can estimate the cost, it shows the price before you pay:
| Where | What you see before paying |
|---|---|
| Video tab | The price per second in the model picker, and the clip's estimated total under the settings and in the confirmation. |
| Photo dropped on the 3D view | The model, the paying account and the estimated price on the card's Make a 3D model row. |
| AI Render | The model, the paying account and the estimated price on the card. |
| Generation Flows | Each paid step's account and estimated price, and the total, in the run's approval. |
| Image and 3D tabs | The model, the account and the uploads. No price is shown here; your provider bills you as usual. |
Prices are estimates from the providers' published price lists, marked ≈. A price Angelo doesn't know reads "price unknown", never zero.
Results and your scene#
Results go to Media Studio's library, shared by all your files, with folders and search. Import adds your own pictures to it. Nothing goes into your scene until you add it. Click a result for its menu:
- Add to Scene: a model stands at the 3D cursor; an image becomes a reference picture there.
- Apply to Selected (images): the selected objects get the picture as their material.
- Add to Chat: attach it to your next message. Use as reference (images): use it for the next generation.
- Export…: save a copy anywhere; an existing file is never overwritten.
Or drag a result into the 3D view, onto an object, the Outliner, a node editor or the chat. A model dragged into the 3D view stands on the surface under the pointer. Adding a result is one step that Undo takes back.
Your models and images are also in Blender's Asset Browser and asset shelves, under Angelo Media Studio. The library lives in ~/Library/Application Support/Angelo/media-studio on a Mac and %LOCALAPPDATA%\Angelo\media-studio on Windows.