Content Ideas

The Content Ideas node turns one or more content recipes and your brand into new post ideas. Each idea borrows the structure of a recipe, that is its format, hook mechanic, beat order and pacing, and never the original's words, footage, music, faces or logos. The node returns a list, so the node after it, such as Generate Script, runs once per idea.

Content Ideas runs on Nodaro Cloud only. It is not available on self-hosted installs.

When to use it

• You have a post that worked, yours or a competitor's, and you want ideas for your brand on the same structure.
• You want several ideas at once, each ready for a script, a shoot or a video node.
• You want to mix what worked in several posts into one batch of ideas.
• You want ideas in a language other than the language of your brand text.

Quick start

Inputs

Several recipes, one run. Connect several Content Recipe nodes, or one Content Recipe that ran once per post. Every recipe goes into the same run of Content Ideas. The node does not run once per recipe.

Outputs

The output, Ideas, is a list with one creative brief per idea. Every connection that leaves it runs the next node once per idea. This works in the editor and in runs started from the API.

• What each run receives. Each run of the next node receives one idea as readable text, with the same fields as below. The brief ends with the post that inspired the idea.
• All ideas at once. Set a connection to Selected to hand the next node every idea as one text, in a single run. See connection modes.

Settings

What each idea contains

What the ideas borrow

Each idea borrows the structure of a recipe: its format, its hook mechanic, its beat order and its pacing. An idea never reuses the original's words, footage, music, faces or logos.

This rule is built into the node, not only asked of the model. Before the model reads a recipe made by Content Recipe, the post's spoken lines, on-screen text and call to action are removed from it, so the model cannot repeat them.

Credits

Content Ideas is charged per batch of up to 5 ideas, by the tier of the AI model:

• Ideas per run. A run of 6 to 10 ideas is billed at the price in the last column, whatever the exact count. 5 ideas on the default model cost 11 credits. 7 ideas on a standard model cost 77 credits. 10 ideas on a premium model cost 110 credits.
• Reasoning effort. From the API, a reasoningEffort of xhigh or max moves the run one tier up, capped at Premium. The settings panel has no reasoning-effort control on this node.
• Failed runs. A run that fails is refunded. See Credits.
• The next node counts once per idea. In the editor, the estimate of a workflow run prices the node after Content Ideas once per idea. With 5 ideas, a Generate Script set to Gemini 3.6 Flash adds 5 × 11 = 55 credits.

What the whole flow costs

A post of up to 60 seconds, from its link to five scripts. Every node keeps its default settings, and Generate Script has Gemini 3.6 Flash chosen under AI Model:

The analysis is most of the price, and it depends on the length of the post. See Video Analysis for its other steps.

Tips

• Write a specific brand text. Name the audience, the tone, an offer and the things the brand never does. The ideas are only as specific as this text.
• Mix recipes from different posts. The ideas then spread across the posts and across hook mechanics.
• Ask for 5 or 10 ideas. Ideas are charged per batch of 5, so 6 ideas cost the same as 10.
• Connect the node, do not name it. In another node's text, {Content Ideas} gives every idea at once, not the current one. Connect the Ideas output instead, so that each run gets one idea. See Prompt variables.

Troubleshooting

The run stops with "connect at least one Content Recipe first". No recipe reaches the Recipes input. Connect a Content Recipe that has a result, and run again. The run stops before anything is charged.

The run ends with "Content ideas failed". The model did not finish the ideas. The credits are refunded. Run the node again.

The next node ran once, with every idea in one text. The connection from Ideas is set to Selected. Set it to Each to run the next node once per idea.

The node says that only the first 10 recipes were used. More than 10 recipes reached the Recipes input. Content Ideas reads the first 10 and ignores the rest.

From the API

POST /v1/content-ideas writes ideas on Nodaro Cloud. The body takes recipes and the optional brand, count, language, llmModel and reasoningEffort.

• recipes is an array of 1 to 20 recipes, each a string: a recipe's text, or its json as a string. The first 10 recipes are read.
• count is the number of ideas, from 1 to 10. The default is 5.

curl -s https://app.nodaro.ai/v1/content-ideas \
  -H "Authorization: Bearer $NODARO_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "recipes": ["<the text output of a Content Recipe job>"],
    "brand": "A small ceramics studio that sells hand-made mugs online. Audience: people who work from home. Tone: warm, never salesy.",
    "count": 5
  }'

The answer is { "jobId": "…" }. Poll the job with GET /v1/jobs/:id/status until its status is completed. Its output_data holds the ideas as json, all of them as one text, and one brief per idea as listResults. See Jobs.

A request that cannot run is refused with 400 before any job exists, so nothing is charged. That happens when no usable recipe is sent, or when the model cannot return structured data. It also happens when a field is out of range: more than 20 recipes, a count outside 1 to 10, a brand over 6,000 characters or a language over 40.

There is no dedicated MCP tool or SDK method. In a workflow, the node's type is content-ideas. AI assistants and code can add it with the workflow tools or the Workflows API.