Hanzo

Prompts

Versioned prompts with labels and history — manage, ship, and roll back prompts through one API.

Prompts

Prompt management with immutable versions: every edit creates a new version, and labels like production or latest point at a specific version. Change what ships by moving a label, not by editing code. Part of the Hanzo Cloud surface, tenant-scoped by org.

Create a prompt

POST /v1/prompts registers a new prompt (or a new version of an existing one). Labels and tags are attached at creation.

curl -X POST https://api.hanzo.ai/v1/prompts \
  -H "Authorization: Bearer hk-..." \
  -H "Content-Type: application/json" \
  -d '{
    "name": "support-reply",
    "type": "text",
    "prompt": "You are a support agent for {{product}}. Answer concisely.",
    "labels": ["production", "latest"],
    "tags": ["support", "routing"]
  }'

Fetch, list, and roll back

List prompt metadata, or fetch one by name to get its full version history. Because versions are immutable, rolling back is just re-pointing the production label at an earlier version.

# List prompts (name, versions, labels, tags)
curl https://api.hanzo.ai/v1/prompts \
  -H "Authorization: Bearer hk-..."

# One prompt, with version history
curl https://api.hanzo.ai/v1/prompts/support-reply \
  -H "Authorization: Bearer hk-..."
{
  "name": "support-reply",
  "type": "text",
  "versions": [1, 2, 3],
  "labels": ["production", "latest"],
  "tags": ["support", "routing"]
}

Metrics

Usage and performance per prompt version — call volume, latency, and cost — are available at GET /v1/prompts/metrics, so you can confirm a new version is an improvement before moving the production label.

  • Experiments — evaluate a prompt version against a dataset
  • Traces — see prompts in the context of live requests
  • LLM Gateway — where prompts are rendered and served

How is this guide?

Last updated on

On this page