Nodaro Docs
DocumentationNode ReferenceModelsAI Agents (MCP)DevelopersSelf-hostingResearch

Legacy nodes

Older workflows can contain node types the node picker no longer offers. See what each one becomes and what changes when the workflow opens.

A legacy node is a node type that an older saved workflow can still contain, but that the node picker no longer offers. When the workflow opens, Nodaro replaces each legacy node with its current node: the settings carry over, and the connections move to the matching inputs. You do not need to rebuild anything, and the workflow runs as it did before.

Which node replaces which

Old nodeAPI typeCurrent node
Text to Videotext-to-videoGenerate Video
Image to Videoimage-to-videoGenerate Video
Image to Imageimage-to-imageModify Image
Edit Imageedit-imageModify Image, Upscale Image or Remove Background
TableloopList
AI Agentai-writerPrompt

The update happens when the workflow opens, so a workflow you have not opened for months still loads and runs. To move a workflow between accounts or installs, see Import and export.

Text to Video and Image to Video

Text to Video made a video from a prompt, and Image to Video animated a still image. Both became one node, Generate Video, which chooses its mode from what you connect: a prompt alone, a start frame, a start and an end frame, or references.

When an old workflow opens:

  • The node becomes a Generate Video node with the same model and settings.
  • The model list, the settings and the credit prices for these uses did not change.
  • The connections move to the new inputs. The old prompt input becomes Prompt. The old reference inputs become Image Refs, Video Refs and Audio Refs. The old cinematography input becomes Look or Elements, depending on the picker that is connected.

Generate Video also adds a text-only mode on models such as Veo 3.1 and Kling, so one node covers both old jobs.

Image to Image

Image to Image changed an existing image with a text prompt. The prompt-driven editing it did is now part of Modify Image. An Image to Image node becomes a Modify Image node when the workflow opens, with its connections unchanged.

Edit Image

Edit Image applied one operation to an image: an upscale, a background removal or a prompt edit. Each operation now has its own node. When the workflow opens, the node becomes the node that matches its operation:

Operation set on the Edit Image nodeBecomes
Nano Banana Edit, a prompt editModify Image
Recraft Remove BG, a background removalRemove Background
An upscale, such as Recraft Upscale or Topaz UpscaleUpscale Image

To chain operations, for example to remove a background and then upscale, connect Remove Background to Upscale Image.

Table

The Table node was a list with several typed columns. It is now part of the List node, which starts as one text column and grows into a table with typed columns when you connect more sources to it.

When an old workflow opens, a Table node becomes a List node. It keeps all its columns and rows, and its connections stay as they were. Published apps that contain a Table node are updated the same way.

AI Agent

The AI Agent node wrote text, such as a set of image prompts, with a language model. It was merged with the text-generation node into one node, Prompt. The Prompt node was called Generate Text until September 2026.

When an old workflow opens:

  • The node becomes a Prompt node and keeps the model it had saved. If no model was saved, it uses Claude Sonnet 4.6, the model the AI Agent node used by default.
  • The connection into its old text input moves to the Prompt input. Its Text output keeps its connections.

Preview

The Preview node shows any result, such as text, an image, a video or audio, inline on the canvas. It comes from a time when nodes did not show their own results.

Preview still works in saved and new workflows. It is in the node picker under Automate › Canvas and Common › Finish. Every node now shows its own result on the canvas, so a Preview node is optional. To gather several results in one place, use Collect. To keep the best of several results, use Choose Best. See Preview.

From the API

The POST /v1/text-to-video, POST /v1/image-to-image and POST /v1/edit-image endpoints still accept requests. For new code, use the endpoints of the current nodes. See Run a single node.

Frequently asked questions

Last updated on

On this page