# NeuralMagic

**Canonical:** https://apis.io/providers/neuralmagic/  
**APIs profiled:** 0

Neural Magic was an AI software company that spun out of MIT in 2018 to build high-performance deep learning inference software. Its flagship open-source products were DeepSparse, a sparsity-aware inference runtime delivering GPU-class performance on commodity CPUs, together with SparseML, SparseZoo, and Sparsify for model compression, sparsification recipes, and a curated model repository, plus nm-vllm, an enterprise LLM inference server for GPUs. Neural Magic was a prolific contributor to the community vLLM project. The company was acquired by Red Hat (agreement announced November 2024, completed January 2025) and its engineering now advances vLLM and Red Hat AI. The community versions of DeepSparse, SparseML, SparseZoo, and Sparsify were deprecated on 2026-06-02. Neural Magic exposed no commercial REST API; its developer surface was a set of first-party Python libraries published to PyPI and open source on GitHub.

## Kin Score — 10.7 / 100 (minimal)

Scored 2026-08-20 under rubric 0.12.0. Trend: flat (+0.0 from 10.7).

| Facet | Score |
|---|---|
| Discoverability | 50.0 |
| Contract Quality | 0.0 |
| Governance | 0.0 |
| Contract Governance | 0.0 |
| Operational Transparency | 18.4 |
| Developer Ergonomics | 16.7 |
| Commercial Clarity | 0.0 |
| Access Clarity | 0.0 |

## Agent readiness — 0.0 (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 | 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

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

## Tags

Company, Artificial Intelligence, Machine-Learning, Inference, LLM, Model Compression, Open-Source, Python, vLLM, Deep Learning

---

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