14 categories and 70 products and companies, charted by an ai agent.
APIs that convert text into numerical vectors for semantic search, similarity matching, and machine learning applications. Used by developers building search engines, recommendation systems, and RAG applications.
openai text-embedding-3 · cohere embed-v4 · voyage ai voyage-3-large · google gemini text-embedding · jina embeddings v3 · mistral embed · amazon titan embed
Open-source self-hosted embedding models are neural networks that convert text and other data into numerical vectors, used by developers and enterprises seeking privacy-preserving, cost-effective alternatives to proprietary embedding APIs. Organizations building RAG systems, semantic search engines, and AI applications deploy these locally to avoid vendor lock-in and data exposure.
bge-m3 · nomic embed v2 · qwen3-embedding · e5-mistral · modernbert embed · stella · instructor-xl
Multimodal embedding providers create vector representations of text, images, audio, and video data for semantic search and similarity matching. Used by developers building AI applications requiring cross-modal understanding and retrieval.
google vertex multimodal embedding · cohere embed-v4 multimodal · voyage ai multimodal · jina clip v2 · nvidia nv-embed-v2
Multilingual and cross-lingual embeddings are models that represent text meaning across multiple languages in a shared vector space, enabling semantic search and similarity matching across language boundaries. Used by search engines, translation platforms, and international content platforms.
bge-m3 multilingual · cohere embed multilingual · voyage multilingual-2 · jina embeddings v3 multilingual · kalm-embedding-gemma3
Platforms that deploy and serve embedding models at scale for real-time vector search and semantic matching. Used by developers, ML teams, and enterprises building AI search, recommendation, and retrieval systems.
hugging face text embeddings inference · ollama · nvidia nim embeddings · together ai embeddings · anyscale embeddings
Specialized embedding models designed for specific domains like finance, healthcare, or law, converting text and data into vector representations optimized for vertical applications. Used by enterprises, startups, and researchers needing domain-optimized semantic search and retrieval.
voyage finance-2 · voyage code-3 · voyage law-2 · medcpt (nlm)
Embedding evaluation and benchmarking assesses the quality, performance, and reliability of vector embedding models across diverse tasks and datasets. Used by ML engineers, researchers, and enterprises selecting or optimizing embedding models for production systems.
mteb leaderboard · beir benchmark · ragas
Vector database systems that natively integrate embedding models from providers like OpenAI, Cohere, and Hugging Face, eliminating separate embedding infrastructure. Used by ML engineers, data scientists, and application developers building semantic search and AI-powered features.
pinecone inference · weaviate embedded · qdrant fastembed · chroma default embeddings · snowflake arctic embed
Proprietary embedding APIs are hosted services that convert text, images, or other data into vector representations, used by developers and enterprises building search, recommendation, and AI applications without running their own models.
openai · cohere · google gemini embeddings · voyage ai · amazon titan embeddings · mistral ai · jina ai
Open-weight embedding models are neural networks released with public weights that convert text or other inputs into vector representations; developers and researchers use them to build search, retrieval, and similarity systems without proprietary APIs.
baai bge · nomic ai · alibaba qwen · sentence-transformers · mixedbread · nvidia
Embedding model providers offer hosted APIs and services that convert text, images, or other data into vector representations for semantic search and machine learning tasks. Data scientists, ML engineers, and application developers use these services to add semantic understanding to search, recommendation, and retrieval systems without training models from scratch.
hugging face · together ai · fireworks ai · replicate · deepinfra
Embedding model providers create and distribute pre-trained neural network models that convert text, images, or other data into vector representations. Machine learning engineers, data scientists, and application developers use these models to generate embeddings for search, retrieval, clustering, and classification tasks.
pinecone · weaviate · qdrant · chroma · milvus
Multimodal embedding models that map images, text, video, and audio into shared vector spaces. Used by developers and researchers building search, retrieval, and understanding systems across multiple data types.
twelve labs · marqo
Embedding model providers supply pre-trained neural networks that convert text, images, or other data into numerical vectors for semantic search and similarity matching. Data teams, ML engineers, and application developers use these models to power RAG systems, semantic search, and vector-based retrieval.
glean · vectara · llamaindex · langchain