System prompt

A system prompt is the persistent base instruction that sets an avatar's role, tone, boundaries, allowed behaviours, and escalation rules across the whole conversation.

How system prompts work

A system prompt is the persistent instruction that defines how a model or agent should behave. It usually sits above the user's messages and gives the avatar its role, boundaries and rules.

In a real-time avatar, the system prompt often carries the persona: what the avatar is allowed to do, how it should speak, which topics it should avoid, and when it should call a tool or hand off to a human.

A concrete example: a healthcare onboarding avatar might be told to explain product setup, avoid giving medical advice, and escalate any account-specific request to an approved workflow.

System prompts are important because they travel through every turn. A weak system prompt leaves behaviour to chance; a strong one gives the avatar consistent guardrails without scripting every possible answer.

What Anam ships

Anam's Cara-4 model delivers expressive real-time avatars with around 150 ms server-side avatar-generation latency once a session is running, across 70+ languages. Builders use JavaScript and Python SDKs or integrations for LiveKit, Pipecat, ElevenLabs Agents, Agora, and VideoSDK. Bring any AI stack including OpenAI, Claude, Gemini, Mistral, Groq, Deepgram, Cartesia, or custom providers. The platform supports WebRTC delivery, SOC 2 Type II, HIPAA, zero data retention, and regional data residency. Sessions stream low-latency audio and video to browsers and native apps.

Explore the real-time avatar API →

Related terms

Persona

A persona is the configured identity of an avatar, including its role, tone of voice, appearance, behaviour, knowledge boundaries, and rules for how it should respond to users.

Prompt

A prompt is the instruction and context given to a language model for a specific response, including the user's message, system rules, retrieved content, and tool results when available.

Hallucination

A hallucination is when a language model gives a confident but incorrect or unsupported answer, which is especially risky in avatar experiences because the response is delivered through a persuasive face and voice.

AI voice agent

An AI voice agent is a speech-first conversational agent that uses speech recognition, an LLM or agent workflow, and text-to-speech to understand spoken input and reply naturally in real time.

Frequently asked questions

What does a system prompt control in an avatar agent?

A system prompt sets the agent's high-priority instructions, including its role, boundaries, tone, tool rules, escalation behavior, and how it should handle uncertainty.

How is a system prompt different from a persona?

A persona describes how the avatar should come across. The system prompt turns that into operating rules the model should follow throughout the session.

What should go into a system prompt for avatars?

Include the avatar's role, target user, speaking style, source-of-truth rules, tool permissions, safety limits, handoff conditions, and what to do when information is missing.

Can users override a system prompt?

They should not be able to. Good agent design treats user messages as lower-priority input and adds protection against prompt injection, unsafe tool use, and unsupported claims.

Last updated: 17th July 2026 · Reviewed quarterly.

← Back to glossary