# Unsloth

**Canonical:** https://apis.io/providers/unsloth/  
**Website:** https://unsloth.ai  
**APIs profiled:** 0

Unsloth is an open-source framework for fine-tuning, reinforcement learning, and running large language models locally. Its Python library trains 500+ models (text, vision, audio, embeddings) roughly 2x faster with up to 70-80% less VRAM via custom Triton kernels and dynamic quantization, and Unsloth Studio provides a local, browser-based GUI for no-code training and chat. Models loaded in Unsloth are exposed through an authenticated, OpenAI-compatible local API endpoint (llama-server) so agents like Claude Code, Codex, and standard OpenAI/Anthropic SDKs can use them; there is no hosted SaaS API. Backed by Lightspeed Venture Partners.

## Kin Score — 26.3 / 100 (thin)

Scored 2026-08-19 under rubric 0.12.0. Trend: flat (+1.1 from 25.2).

| Facet | Score |
|---|---|
| Discoverability | 57.4 |
| Contract Quality | 0.0 |
| Governance | 4.5 |
| Contract Governance | 4.5 |
| Operational Transparency | 21.1 |
| Developer Ergonomics | 54.8 |
| Commercial Clarity | 31.6 |
| Access Clarity | 31.6 |

## Agent readiness — 8.5 (agent-aware)

| Dimension | Value |
|---|---|
| Spec Presence | no |
| Agentic Access | no |
| Reversibility Documented | no |
| MCP Server | no |
| Auth Clarity | yes |
| Idempotency | no |
| Error Semantics | no |
| OpenAPI Examples | no |
| 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

Self-serve signup — onboarding: self-serve, pricing: unknown, trial: no (confidence: medium).

## Security (2)

- **Unsloth Authentication** — apiKey · 1 scheme
- **Unsloth Domain Security** — TLSv1.3 · HSTS · DMARC

## Tags

Company, Artificial Intelligence, LLM Fine-Tuning, Machine Learning, Open Source, Reinforcement Learning, Quantization, Local AI, Developer Tools

---

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