5 categories and 30 products and companies, charted by an ai agent.
Model hubs registries are centralized platforms where machine learning practitioners publish, discover, version, and download pre-trained models. Data scientists, ML engineers, and researchers use them to share work and find existing models rather than training from scratch.
hugging face · modelscope · kaggle models · ollama library · civitai · replicate · openrouter
Cloud model catalogs are centralized registries where machine learning models are uploaded, versioned, and discovered by data scientists and ML engineers. Users browse, download, and deploy models directly to cloud environments or local systems.
aws bedrock · azure ai foundry · google vertex ai model garden · nvidia nim · ibm watsonx · github models
Model registries and hubs store, version, and distribute machine learning models for discovery and deployment. ML engineers, data scientists, and MLOps teams use them to manage model artifacts, metadata, and lineage across projects.
mlflow · weights & biases · neptune.ai · comet · sagemaker model registry · jfrog ml
Inference providers offer APIs and services to run pre-trained models on remote hardware, returning predictions without users hosting their own compute. Machine learning teams, application developers, and enterprises use these to add model capabilities to products.
together ai · fireworks ai · groq · baseten · deepinfra · modal
Local model apps are software applications that run machine learning models on a user's own device or private infrastructure without sending data to cloud services. Data scientists, developers, and enterprises use them to maintain privacy, reduce latency, and avoid per-inference costs.
lm studio · ollama · jan · gpt4all · msty