Vector database

A vector database stores embeddings for semantic search, so an avatar or agent can retrieve relevant knowledge base content for RAG and answer user questions with the right context.

How vector databases work

A vector database stores embeddings: numerical representations of text, images or other data. Similar items sit close together in vector space, which makes semantic search possible.

In an avatar system, a vector database is often part of RAG. The user's question is converted into an embedding, the database finds relevant chunks of knowledge, and the model uses those chunks to answer.

A concrete example: a user asks, "Can this run in the EU?" The system retrieves data residency and deployment-region notes, then the avatar answers from that source material.

The value is not the database by itself. It is the ability to retrieve the right context quickly enough that the avatar can stay grounded without making the conversation feel slow.

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

RAG

RAG, or retrieval-augmented generation, is a pattern where an agent retrieves relevant documents or data at query time and uses them to produce a more accurate, grounded response.

Grounding

Grounding is the practice of anchoring a model's answer to trusted source content, retrieved documents, or tool results so an avatar can respond with verified information instead of guessing.

Function calling

Function calling lets an LLM invoke approved tools or APIs during a conversation, so an avatar can fetch data, perform actions, or answer from live systems instead of relying on text generation alone.

Customer support agent

A customer support agent is an interactive avatar configured to answer customer questions, retrieve approved help content, check account context, and route complex issues to the right workflow.

Frequently asked questions

What is a vector database used for in avatar agents?

A vector database stores embedded chunks of content so an avatar agent can search by meaning and retrieve relevant context before answering a user.

Why use vector search instead of keyword search?

Vector search can match similar meaning even when the user uses different words from the source document. Many production systems combine it with metadata and keyword filters.

What content belongs in a vector database?

Use approved material the avatar may answer from, such as documentation, help articles, policies, API references, product notes, support content, and internal knowledge bases.

What makes a vector database useful for real-time avatars?

It needs fast retrieval, fresh content, good chunking, metadata filters, permission controls, and relevance testing so answers are both accurate and quick enough for a live session.

Last updated: 17th July 2026 · Reviewed quarterly.

← Back to glossary