Using Media Context in Fluidfit

Add your media and files once, then reuse them across your flow without repeatedly attaching or connecting them to individual nodes.

Media Context lets you save images, videos, audio files, and documents in one place and reference them wherever needed. You can also save reusable text inputs alongside your media.

This makes flow creation faster, reduces repetitive connections, and helps keep your workflows clean and easy to manage.

What can you add?

Media Context supports the following input types:

Input type	Examples
Images	Product images, brand logos, reference images, and design assets
Videos	Product videos, reference footage, and campaign videos
Audio	Voice recordings, music, and audio references
Files	Documents and other supported files
Text	Brand names, product descriptions, instructions, and other reusable text

How to add media and text inputs

Open Flow Context in your flow.
Click Add Inputs.
Select Add Media to upload an image, video, audio file, or document. Select Add Text to create a reusable text input.
Give the input a clear, unique name.
Upload your media or enter your text, then click Save.

Your inputs will appear in Flow Context, where you can search, filter, and manage them.

How to use Context inputs in prompts

You can reference saved inputs in LLM prompts and generative prompt editors. However, the syntax depends on how you want to use the input.

1. Using media in LLM prompts

When using an LLM node, refer to the saved media directly by its input name using double curly braces.

Example:

Analyze this image {{Product_Image}} and create a product description for {{Brand}}.

The LLM can use the referenced media as context while processing your prompt. You don't need to attach the same image or file separately to the LLM node.

You can use multiple media and text inputs in the same prompt.

2. Using media in generative prompts

Generative prompt editors use a slot-based mechanism to associate the right media with the right input field.

Use # followed by the field name, then assign the saved input using double curly braces.

Examples:

What you want to use	Syntax
First Frame	`#firstframe={{Product_Image}}`
Last Frame	`#lastframe={{Ending_Image}}`
Reference Images	`#reference_images={{Image_1}},{{Image_2}}`
Reference Video	`#reference_video={{Campaign_Video}}`
Text input	`#text={{Product_Description}}`

The available fields depend on the selected tool or model. Use the supported field names shown for that generation workflow.

Important: Simply writing {{Product_Image}} in a generative prompt does not assign it to a specific media slot. Use the appropriate #field= syntax to tell the model where the input should be used.

3. Using text inputs

Text inputs can be referenced in two ways, depending on how you want to use them.

As a prompt variable

Use double curly braces directly:

Create an advertisement for {{Product_Name}} using a {{Brand_Tone}} tone.

As a Text input element

If you want to assign the saved text to a specific Text input field in a generative prompt editor, use:

#text={{Product_Description}}

This distinction helps ensure the text is used in the intended way.

Managing your inputs

You can manage all saved inputs directly from Flow Context.

Search: Find an input by its name.
Filter: View All, Images, Videos, Text, or Files.
Edit text: Update the saved text value.
Edit media: Rename an input or replace its uploaded file.
Download: Download an uploaded media file where available.

When you rename an input, its existing references in that flow's prompts are updated automatically. You can also replace a media file while keeping the same input name and references.

A few things to keep in mind

Use clear and unique names for your inputs so they're easy to identify and reuse.
Input names should be referenced exactly as they appear in Flow Context.
For LLM prompts, media can generally be referenced directly using {{input_name}}.
For generative prompts, use the appropriate #field={{input_name}} syntax to assign media or text to the intended input field.
Supported media types and prompt fields may vary depending on the selected model or tool.
You don't need to attach or connect the same saved media separately to every supported node.

In short

Add your media and text once, give each input a name, and reuse it throughout your flow. Reference media directly in LLM prompts or assign it to the right field in generative prompts—all without repeatedly uploading files or creating unnecessary connections.

Flow Context