# MosaicML (Databricks Mosaic Research)

> MosaicML is a machine learning systems company that builds open-source tooling for training and deploying large neural networks and generative AI models faster and at lower cost. Its managed training/inference platform is now offered as Databricks Mosaic AI following the 2023 Databricks acquisition. The developer surface is Python packages, the mcli command-line tool, and open-source GitHub repositories - there is no standalone public REST/OpenAPI.

## Docs
- [Documentation](https://docs.mosaicml.com/en/latest/): Project documentation hub
- [Composer getting started](https://docs.mosaicml.com/projects/composer/en/latest/getting_started/installation.html): Install and first training run
- [mcli CLI docs](https://docs.mosaicml.com/projects/mcli/en/latest/): Command-line interface for the training/inference platform
- [Databricks Mosaic Research](https://www.databricks.com/research/mosaic): Company / research blog

## Packages
- [mosaicml (Composer)](https://pypi.org/project/mosaicml/): PyTorch training speedup library - pip install mosaicml
- [llm-foundry](https://pypi.org/project/llm-foundry/): Train, finetune, evaluate and deploy LLMs - pip install llm-foundry
- [mosaicml-streaming](https://pypi.org/project/mosaicml-streaming/): Stream datasets from cloud object stores - pip install mosaicml-streaming
- [mosaicml-cli (mcli)](https://pypi.org/project/mosaicml-cli/): Platform CLI - pip install mosaicml-cli

## Source
- [Composer](https://github.com/mosaicml/composer)
- [LLM Foundry](https://github.com/mosaicml/llm-foundry)
- [Streaming](https://github.com/mosaicml/streaming)
- [GitHub organization](https://github.com/mosaicml)

## Note
- No public REST/OpenAPI, AsyncAPI, MCP server, OAuth scopes, or /.well-known/ discovery documents are published. This profile is generated by the API Evangelist enrichment pipeline from the public developer surface.