Embedding Providers

Qdrant stores and searches vectors from any embedding model. Each provider page covers setup, model dimensions, document and query embedding, and the requirements specific to that provider.

Generate Embeddings with Qdrant

Managed

Qdrant Cloud Inference

Embed text and images on Qdrant Cloud with hosted dense, sparse, and multivector models, including free models. Connect OpenAI, Cohere, Jina AI, and OpenRouter.

Use Cloud Inference

Local

FastEmbed

Run open dense, sparse, and multivector models on your own hardware with Qdrant's lightweight Python library. You do not need an API key or GPU.

Browse Supported Models

Choose an inference approach to compare generating embeddings in your application, in the Qdrant cluster, and on Qdrant Cloud.

Explore All Providers

16 providers

Qdrant Cloud Inference

Embedding models hosted by Qdrant Cloud, plus external providers, called through the Qdrant API with no embedding service to run.

Input: Text, Image
Vectors: Dense, Sparse, Multivector
Runs On: Qdrant Cloud Inference

FastEmbed

Qdrant's Python library for running dense, sparse, image, and multivector embedding models on your own hardware.

Input: Text, Image
Vectors: Dense, Sparse, Multivector
Runs On: Self-hosted

Amazon Bedrock

Amazon Titan, Amazon Nova, Cohere, and TwelveLabs embedding models through Amazon Bedrock, in your AWS account.

Input: Text, Image, Video, Audio
Vectors: Dense
Runs On: Provider API

Cohere

Multimodal Embed models for text and images, with separate document and query input types.

Input: Text, Image
Vectors: Dense
Runs On: Provider API, Qdrant Cloud Inference

Gemini

Multimodal embedding model for text, images, video, audio, and PDFs, with adjustable dimensions.

Input: Text, Image, Video, Audio, PDF
Vectors: Dense
Runs On: Provider API

Jina AI

Multilingual text, code, and multimodal embeddings with task adapters and multivector output.

Input: Text, Image, Video, Audio, PDF
Vectors: Dense, Multivector
Runs On: Provider API, Qdrant Cloud Inference

Mistral

Text and code embedding models through the Mistral API, or through OpenRouter on Qdrant Cloud.

Input: Text
Vectors: Dense
Runs On: Provider API, Qdrant Cloud Inference

Mixedbread

Open-weight text embedding models to run yourself, on Qdrant Cloud Inference, or through a legacy API.

Input: Text
Vectors: Dense
Runs On: Provider API, Self-hosted, Qdrant Cloud Inference

Nomic

Open text and image embedding models with Matryoshka sizes, through the Nomic API or FastEmbed.

Input: Text, Image
Vectors: Dense
Runs On: Provider API, Self-hosted

NVIDIA

NVIDIA Nemotron text and image embedding models, through the NVIDIA API or self-hosted NIM containers.

Input: Text, Image
Vectors: Dense
Runs On: Provider API, Self-hosted

Ollama

Runs open embedding models such as Qwen3 Embedding and EmbeddingGemma 2 on your own hardware.

Input: Text
Vectors: Dense
Runs On: Self-hosted

OpenAI

Text embedding models with adjustable dimensions, through the OpenAI API or Qdrant Cloud Inference.

Input: Text
Vectors: Dense
Runs On: Provider API, Qdrant Cloud Inference

Superlinked

Open-source inference server for 100+ dense, sparse, and multivector models, self-hosted or serverless.

Input: Text
Vectors: Dense, Sparse, Multivector
Runs On: Self-hosted, Provider API

TwelveLabs

Marengo models that embed video, audio, images, text, and documents in one vector space.

Input: Text, Image, Video, Audio, PDF
Vectors: Dense
Runs On: Provider API

Upstage

Solar text embeddings with paired passage and query models, through an OpenAI-compatible API.

Input: Text
Vectors: Dense
Runs On: Provider API

Voyage AI

Text, code, and multimodal embedding models, plus the open-weight voyage-4-nano for local use.

Input: Text, Image, Video
Vectors: Dense
Runs On: Provider API, Self-hosted

Was this page useful?

On this page:

Ready to try Qdrant?