# MosaicML

**Canonical:** https://apis.io/providers/mosaicml/  
**Website:** https://www.databricks.com/research/mosaic  
**APIs profiled:** 0

MosaicML was a San Francisco-based foundation model training company founded in 2021 by Naveen Rao and Hanlin Tang to make large-scale model training faster and cheaper through algorithmic and systems efficiency. Databricks acquired MosaicML in July 2023 for approximately $1.3 billion and folded the team and platform into Databricks Mosaic AI Research. The MosaicML training platform is now delivered as Databricks Mosaic AI Training (pretraining and finetuning), and the original mosaicml.com domain redirects to the Databricks Mosaic Research microsite. MosaicML's surviving open-source artifacts include Composer (a PyTorch training library with built-in speedup recipes), Streaming (a cloud-native dataset format for efficient distributed training), LLM Foundry (the training code behind Databricks' DBRX foundation model), the Diffusion training stack, and the MCLI command line and Python SDK that orchestrate Pretraining and Finetuning jobs against the managed training service. The Mosaic AI Training service itself is accessed exclusively through the MCLI command line, the mosaicml-cli Python SDK, and Databricks workspace integrations — there is no publicly published REST OpenAPI surface for the training control plane. Commercial access requires a Databricks account; pricing is consumption-based via Databricks DBU billing rather than an independent MosaicML pricing page.

## Kin Score — 11.5 / 100 (emerging)

Scored 2026-08-20 under rubric 0.12.0. Trend: flat (+0.0 from 11.5).

| Facet | Score |
|---|---|
| Discoverability | 50.0 |
| Contract Quality | 1.4 |
| Governance | 0.0 |
| Contract Governance | 0.0 |
| Operational Transparency | 5.3 |
| Developer Ergonomics | 16.7 |
| Commercial Clarity | 10.5 |
| Access Clarity | 10.5 |

## Agent readiness — 1.5 (human-only)

| Dimension | Value |
|---|---|
| Spec Presence | no |
| Agentic Access | no |
| Reversibility Documented | no |
| MCP Server | no |
| Auth Clarity | no |
| Idempotency | no |
| Error Semantics | no |
| OpenAPI Examples | documented |
| Rate Limit Signal | no |
| Event Surface Described | no |
| Agent Skills | no |
| Well Known Catalog | no |
| Consent Identity | no |
| Agent Card | no |
| Dry Run Mode | no |

## Access

Unknown — onboarding: unknown, pricing: unknown, trial: no (confidence: low).

## Security (2)

- **Mosaicml Domain Security** — TLSv1.3 · HSTS · DNSSEC · DMARC
- **Mosaicml Vulnerability Disclosure** — security.txt · contact published

## Tags

Artificial Intelligence, Foundation Models, Model Training, Pretraining, Fine-Tuning, LLM, Generative AI, PyTorch, Distributed Training, GPU, Databricks, DBRX, Composer, Streaming, LLM Foundry

---

Profiled by [API Evangelist](https://apievangelist.com) and published on [APIs.io](https://apis.io/providers/mosaicml/). Scores are computed from the provider's own public artifacts under a published rubric.