An Embedding Model is an AI model that converts text, images, or other data into a list of numbers called a vector.

These numbers represent the meaning and characteristics of the input data in a mathematical form that computers can compare.

1. Simple Example

Suppose we have three sentences:

• “How to reset my password?”
• “I forgot my login credentials.”
• “How to transfer money?”

An embedding model converts these sentences into vectors.

Illustrative vectors only (real embeddings may contain hundreds or thousands of dimensions).

Text	Vector
Reset my password	[0.21, 0.85, 0.32, ...]
Forgot login credentials	[0.24, 0.82, 0.35, ...]
Transfer money	[0.91, 0.12, 0.67, ...]

The first two vectors are likely to be closer to each other because their meanings are similar.

The third vector represents a different concept.

This allows AI systems to search based on meaning rather than just matching words.

2. How Embedding Models Work with Vector Databases

3. Popular Embedding Models

Model	Provider
text-embedding-3-small	OpenAI
text-embedding-3-large	OpenAI
nomic-embed-text	Nomic / Ollama
all-MiniLM-L6-v2	Sentence Transformers
bge-m3	BAAI

Some models can run locally, while others are available through APIs.

4. Embedding Model vs LLM

Embedding Model	LLM
Converts text into vectors	Generates text and answers
Used for semantic search	Used for reasoning and generation
Numerical output	Text output (typically)
Example: nomic-embed-text	Example: llama3

5. Example in a Local AI Application

If you use Ollama:

ollama pull nomic-embed-text

Generate embeddings:

curl http://localhost:11434/api/embed \
  -d '{
    "model": "nomic-embed-text",
    "input": "How to reset banking password?"
  }'

The response contains a vector that you can store in PostgreSQL with pgvector or another vector database.

The complete AI knowledge-search flow

Document → Embedding Model → Vector Database → Similarity Search → LLM → Answer

The important distinction:

• Embedding model understands the semantic representation of data.
• Vector database stores and retrieves that representation.
• LLM uses retrieved information to answer questions.