Start with a pre-built wrapper code instead of a blank prompt. Pass basic fields or files, run your selected model, and download the generated artifact.
Recommended start path
Pick a real use case first, then copy the wrapper-code API request.
Quickstart
Authenticate with x-api-key or Bearer zw_live_... from your dashboard.
Call reusable prompt behavior by code instead of writing prompts from scratch.
Send variables, plain text, or uploaded files as context for the selected wrapper.
Every execution records model, tokens, cost, artifact, billing source, and trace ID.
Examples
These examples show the wrapper-code API contract. Keep idempotency keys unique per operation to avoid duplicate executions.
curl -X POST https://api.zywrap.com/v2/runs \
-H "x-api-key: $ZYWRAP_API_KEY" \
-H "Content-Type: application/json" \
-H "Idempotency-Key: first-run-001" \
-d '{
"model": "openai-gpt-4o-mini",
"wrapperCodes": ["mc_cold_outreach_email_b2b_base"],
"input": {
"text": "Write a concise outbound email for an AI workflow API used by SaaS developers."
},
"output": {
"format": "markdown",
"artifact": true
}
}'Choose a category, use case, wrapper variation, model, file inputs, and overrides. Then run sync or async and download the artifact.