Category
Hosted platforms for accessing foundation models, training, fine-tuning, evaluation, and inference APIs.
21 tools in this category
Microsoft
Azure-hosted access to OpenAI models with enterprise controls, networking, and governance.
Azure OpenAI Service is used when teams want to build AI applications with managed model endpoints while keeping deployment inside Azure's enterprise security, compliance, and networking boundaries.
Hugging Face
A repository and platform for AI models, datasets, demos, and hosted inference options.
Hugging Face Hub is used to discover models, share datasets, run demos, host spaces, and connect development workflows to the broader open AI ecosystem.
OpenAI
A hosted API platform for OpenAI models, assistants, realtime experiences, and AI application building.
OpenAI Platform is used to add language, vision, audio, and realtime AI features to applications through APIs rather than using a consumer chat product directly.
Amazon Web Services
An AWS-managed platform for foundation model APIs, agents, evaluation, and enterprise AI services.
Amazon Bedrock is used to access multiple model providers through AWS, build generative AI applications, and keep workloads close to existing AWS security and infrastructure controls.
Anthropic
Anthropic's API platform for Claude models and enterprise AI application development.
Anthropic's API is used to embed Claude models into products, automate reasoning-heavy tasks, and build business applications that need a hosted model endpoint.
Cloudflare
A serverless inference platform that lets developers run AI models close to end users on Cloudflare's network.
Workers AI is used to add generation, embeddings, translation, and media AI capabilities to applications with edge deployment patterns.
Cohere
A model and enterprise AI platform for language, search, reranking, and generative use cases.
Cohere Platform is used to power enterprise search, retrieval, generation, reranking, and custom AI applications that need hosted model access and business-oriented tooling.
Databricks
Databricks' model and AI application stack for training, deployment, evaluation, and governed production usage.
Mosaic AI is used to build retrieval systems, fine-tune or serve models, evaluate AI applications, and operationalize generative workloads on data lakehouse infrastructure.
DataRobot
An enterprise AI platform for model development, governance, and operational deployment.
DataRobot AI Platform is used by organizations that need managed tooling for model building, MLOps, governance, and broader enterprise AI operations.
Exa
A search infrastructure company offering AI-native web search and retrieval tools for developers and research apps.
Exa is used to power AI search experiences, retrieve high-quality web results, and build research or agent products with search APIs.
Fireworks AI
A model-serving platform focused on fast inference and API access for open and proprietary models.
Fireworks AI is used to serve models through hosted endpoints, optimize latency-sensitive inference, and give developers API access to a broad model catalog.
GitHub
A developer-facing GitHub feature for experimenting with and comparing hosted AI models.
GitHub Models is used to test prompts, compare model outputs, and integrate model access into GitHub-centric development workflows.
A browser-based environment for prototyping with Gemini models and model APIs.
Google AI Studio is used to test prompts, inspect model responses, and prototype Gemini-powered applications before moving into a more structured production environment.
Groq
A hosted inference platform focused on low-latency model serving.
GroqCloud is used when developers want very fast inference for chat, speech, or agent workloads and prefer a hosted platform instead of operating custom performance hardware.
IBM's enterprise platform for foundation model access, tuning, governance, and AI application development.
watsonx.ai is used to build and govern enterprise AI applications, access hosted models, and run generative workloads under IBM's platform controls.
Mistral AI
Mistral's hosted platform for model APIs, fine-tuning, and enterprise deployment options.
La Plateforme is used to call Mistral models through hosted APIs, evaluate model behavior, and integrate European-focused model infrastructure into applications.
OpenRouter
A unified API layer for routing requests across many hosted model providers.
OpenRouter is used to compare models, switch providers quickly, and simplify app integration when one product needs access to many model backends.
Replicate
A hosted platform for running and integrating AI models through simple APIs.
Replicate is used to add hosted model inference to products quickly, especially for image, video, and open-model experiments that benefit from API access without managing GPUs directly.
Snowflake
Snowflake's AI layer for model access, document processing, and in-platform generation over enterprise data.
Snowflake Cortex AI is used to build data-aware assistants, run inference close to governed enterprise data, and extend analytics workflows with generation.
Together AI
A cloud platform for model inference, fine-tuning, and open model access through APIs.
Together AI is used to deploy and consume open models, fine-tune workloads, and build AI products on top of hosted inference infrastructure.
Google Cloud's AI platform for managed models, training, agent tooling, and generative AI workloads.
Vertex AI is used to build production AI systems that need managed model access, training pipelines, evaluation, retrieval, and enterprise deployment inside Google Cloud.