Nodes
Image node
Generate images from a prompt, put screenshots, diagrams and photos on the canvas so chats can see them, and keep every image next to the work it came from.
Images are context too. Put a screenshot of an error, a chart from a paper or a whiteboard photo on the canvas, connect it to a chat, and the model looks at the picture itself, next to your notes and sources. An empty Image node is also a prompt box, so you can generate a picture where you want it on the canvas.
Add an image
- Toolbar: press
U(or click Upload) and pick a file. Each image becomes its own node. - In an empty Image node: press
I(or click Image node in the toolbar), click the canvas, then click Upload an image in the node. Right-clicking the canvas and choosing Image Node works too. - Drag and drop: drop image files onto the canvas. Each file becomes its own node.
- Paste: copy an image or a screenshot, then press
⌘V(Ctrl+Von Windows and Linux) on the canvas.
Upload an image takes PNG, JPG, GIF and WebP files up to 50 MB; drag and drop also accepts SVG. The title starts as the file name; click it to rename.
View and find the file
Hover over the node for Reveal in Folder, Duplicate, Delete and Open in sidebar (the side view). To see an image larger, zoom the canvas or drag the node's left, right or bottom edge.
| How you added it | Where the file is |
|---|---|
| Upload, drag and drop | Where it already was. The node points at the original, so moving or deleting that file breaks the node. |
| Pasted (screenshots, images copied from a browser) | ~/SlashspaceOS/images/pasted |
| Generated | An images folder inside the canvas's folder |
Use an image as context
Connect an image you added to a chat like any other source: drag from the image's right handle to the chat, or type @ in the chat's composer and pick the image by its title. The chat sends the image with each message. See Connections and context.
The chat's model has to accept images. Auto picks an image-capable model when an image is connected, but some models are text-only (for example GPT-OSS 120B) and Slashspace doesn't warn you. If a model ignores the picture or returns an error, switch models.
Note: An image reaches the chat only if it is 20 MB or smaller and stored inside your home folder. Images on an external drive aren't sent.
Generate an image
An empty Image node is a prompt box:
- Press
I(or click Image node in the toolbar) and click the canvas. - Type what you want in "Describe an image".
- Pick a model and an aspect ratio.
- Press
Enteror click Generate.
The node shows the seconds so far with a Stop button, then the image. Your prompt goes straight to the model, with no chat turn in front of it.
The model dropdown lists your own keys first, then Slashspace's models:
- Your own OpenAI, Google, fal or xAI key. The image is made straight from your computer, and that provider bills you. Save a key in API Providers.
- Slashspace's hosted models. These need no provider account. They're paid from your optional hosted wallet, and the dropdown shows about what each one costs. Stop only stops the waiting: a run that already started is still charged.
Use references
Connect Image and text nodes into an Image node, and they appear as chips above the prompt. Images become references, and the box then reads "Describe a change". Text goes in front of your prompt. Click a chip to leave it out of the next run.
Edit and regenerate
Hover a finished image for these buttons, or press the shortcut with the node selected.
- Edit (
E): opens a new connected node with this image as its reference, so the original stays as it is. This works on any image, generated or added. - Regenerate (
R): on an image made in this node, makes a new one in place from the same prompt. Undo brings the last one back.
If a generation fails
The reason shows in the node.
| Message | What to do |
|---|---|
| "Not enough Slashspace credits for this model." | Add funds to the wallet, or click Use your own key. |
| "Your OpenAI key was rejected. Check it in Settings." | Click Open settings and check the key. Other providers show the same message with their name. |
| "OpenAI needs your organization verified before it can make images." | Verify your organization with OpenAI, or pick another model. |
| "Google image models need billing turned on for your key." | Turn on billing with Google, or pick another model. |
| "You're offline. Connect to generate." | Reconnect. Every model needs the internet, even with your own key. |
| "Image generation comes with Slashspace Yearly and Lifetime." | Shown on a greyed-out Slashspace model when your plan doesn't include it. See Plans and billing. |
Images from a chat
You can also generate from a chat. In a chat set to Image mode, each generation branches into its own connected Image node titled "Generated image", with the prompt underneath. Image generation covers prompts, models and cost.
| What the node shows | What it means |
|---|---|
| "Waiting to start…" or "Generating image…" | On its way. Once it's generating, Cancel stops it. |
| Before I generate | The model needs a decision. Click an option, or type in "Or tell me what to change…" and press Enter. |
| The finished image | Edit opens a new Image-mode chat connected to it. The clipboard button copies it. "via your OpenRouter key" means your own key paid. |
| Upgrade to generate images | Your plan doesn't include hosted image generation. See Plans and billing. |
| Out of usage balance | Your hosted wallet is empty. Add funds opens your account in the browser. |
| Image generation cancelled or Image generation failed | You stopped it, or the line underneath says what went wrong. Adjust the prompt in the chat and send again. |
To build on any image, generated or added, connect it into an Image node or an Image-mode chat as a reference for the next generation.
Start a new node from an image
Drag from the right handle and let go on empty canvas. A Create node menu offers Chat (a new chat with the image already connected), Image (an empty Image node) and Scratch pad (a new Scratch pad). Press Esc to close it.