Mistral AI

Mistral AI builds configurable, high-performance large language models and a unified platform to run, fine-tune, and deploy them across cloud, on-premises, edge, and devices. The company provides APIs and developer SDKs for chat, embeddings, classifications, agents, and multimodal features, plus tooling for privacy-preserving private deployments and enterprise orchestration. Mistral emphasizes open-source models, expert-led support, and production-ready integrations for teams and organizations.

• State-of-the-art models — Access Mistral’s family of open-source and commercial models via the API or SDK; choose model size and latency trade-offs to match your deployment needs and scale.
• APIs for chat, embeddings, and multimodal — Use the documented REST API endpoints for chat completions, embeddings, OCR, audio transcription, classifiers, and batch processing to integrate AI into applications.
• Agents and orchestration — Build and deploy agentic workflows with the platform’s agent endpoints and orchestration primitives to connect tools, web searches, and connectors.
• Fine-tuning and custom models — Fine-tune or distill models with your data and deploy private variants for improved domain performance and data control.
• Private deployments — Run models on cloud, on-premises, edge, or device environments with enterprise tooling for SSO, audit logs, and data sovereignty.
• Developer SDKs and examples — Start with official SDKs and code samples to call chat completions, embeddings, and agent endpoints; adjust temperature, max tokens, and tool integrations to tune behavior.

To get started, sign up for an API key, pick a model, and call the chat or embeddings endpoints; for agent workflows, follow the Agents section in the API docs to register tools and enable orchestration. For private deployments, contact sales for enterprise offerings and deployment guidance.