The system prompt is the most important configuration for your agent. It defines how the agent thinks, what it knows, what it avoids, and how it responds to users.
The agent_prompt Field
Set the agent's system prompt using the agent_prompt field in the update-agent endpoint. Use the Authorization: Token prefix for API keys.
cURL
Prompt Best Practices
Define the Role
Start with "You are a [role] for [company]" to anchor identity.
Set Boundaries
Specify what the agent should NOT discuss to prevent off-topic loops.
Escalation Paths
Tell the agent when to offer human intervention.
Plain Text Only
Vozzo optimizes for plain text instructions without HTML/Markdown.
AI Auto-Edit Prompt
Let Vozzo's AI suggest improvements to your prompt based on your feedback. Note the use of issue_description in the request body.
cURL
Sub-Prompts & Snippets
Sub-prompts are reusable prompt snippets that get appended to the main prompt.
cURL
Prompt History
Every update creates a new version. Retrieve history to track changes or audit outputs.
cURL
| Parameter | Type | Description |
|---|---|---|
page | integer | Page number (default: 1) |
size | integer | Items per page (default: 10) |
camp_template | boolean | Filter campaign-specific prompts |
Prompt Fields Reference
| Field | Type | Description |
|---|---|---|
agent_prompt | string | Main system prompt (plain text) |
agent_behaviour | string | Behavioural tone modifier |
sub_prompt_data | integer[] | IDs of sub-prompt snippets to append |
fallback_config_json | array | Configuration for fallback scenarios |

