Introduction to Embed

The circuit.core.embed primitive lets you call an existing Circuit (yours or a community-shared one) as a sub-workflow within your own Circuit. It:

1. Posts your inputs and target CircuitVersion to the ModelWorks runtime
2. Waits for that embedded Circuit to finish
3. Pulls back all of its outputs (and performance reports) into your parent Circuit’s execution context

Use embed to compose higher-level logic, share reusable building blocks, and foster ecosystem growth.

Step Syntax

{
  "name": "<step_name>",
  "description": "Optional description",
  "function": "circuit.core.embed",
  "conditions": [
    { "expression": "${{ <selector> }}" }
  ],
  "query": {
    "context": {
      "version": "<TARGET_VERSION_UID>"
    },
    "input": {
      /* key/value pairs matching the embedded Circuit’s inputs */
      "foo": "${{ input.foo }}",
      "bar": 123
    }
  }
}

• name A unique identifier for this step.

• description (optional) Document purpose or behavior.

• function Must be "circuit.core.embed".

• conditions (optional) Array of pre-conditions (selectors) that must all resolve to true before running.

• query.context.version The UUID of the CircuitVersion to invoke.

• query.input An object whose keys exactly match the target Circuit’s input names. Values can be literals or selectors like "${{ ... }}".

Outputs & Performance

When the embedded Circuit completes:

• All of its outputs appear under <step_name>.output.outputs.<outputName>
• Performance reports (time taken, CPU/RAM usage) appear under <step_name>.output.performance_reports

Reference them via JSON-style selectors:

"${{ openai_query_extend.output.outputs.chat_completion_output }}"

Cost Implications

Embedding a Circuit adds that Circuit’s credit cost to your parent Circuit’s total. This:

• Keeps billing transparent
• Rewards authors for shared Circuits
• Enables rapid composition without duplicating logic

Example

Embed a shared “OpenAI-4.0” chat Circuit:

{
  "name": "openai_query_extend",
  "description": "Invoke community Chat-4.0 Circuit",
  "function": "circuit.core.embed",
  "conditions": [
    { "expression": "${{ input.messages.length > `0` }}" }
  ],
  "query": {
    "context": {
      "version": "01e03926-ad5f-4992-9b18-56bbfb192c72"
    },
    "input": {
      "messages": "${{ input.messages }}"
    }
  }
}

Then consume its output:

{
  "name": "final_output",
  "value": "${{ openai_query_extend.output.outputs.chat_completion_output }}"
}

By embedding optimized, battle-tested Circuits you accelerate your own workflows and help grow the ModelWorks community!