# Refuel

**Canonical:** https://apis.io/providers/refuel-ai/  
**Website:** https://www.refuel.ai  
**APIs profiled:** 3

Refuel is an AI data-labeling and data-enrichment platform that uses large language models to label, clean, structure and enrich enterprise datasets. Refuel Cloud exposes a REST API at cloud-api.refuel.ai covering projects, datasets, tasks and task runs, taxonomies, seedsets and evalsets, confidence calibration, Refuel LLM-2 finetuning, and deployed applications whose realtime label endpoint transforms new rows on demand. The open-source autolabel library lets teams run the same LLM labeling workflows in their own environment against OpenAI, Anthropic, Google, HuggingFace, vLLM or Refuel-hosted models. Refuel.ai was acquired by Together AI in May 2025; the platform continues to operate and its API is still live.

## Kin Score — 64.6 / 100 (strong)

Scored 2026-08-25 under rubric 0.14.0. Trend: flat (+0.0 from 64.6).

| Facet | Score |
|---|---|
| Discoverability | 81.5 |
| Contract Quality | 66.3 |
| Governance | 30.3 |
| Contract Governance | 30.3 |
| Operational Transparency | 34.2 |
| Developer Ergonomics | 69.6 |
| Commercial Clarity | 89.5 |
| Access Clarity | 89.5 |

## Agent readiness — 41.8 (agent-ready)

| Dimension | Value |
|---|---|
| Spec Presence | yes |
| Agentic Access | derived |
| Reversibility Documented | no |
| MCP Server | documented |
| Auth Clarity | bearer |
| Idempotency | no |
| Error Semantics | verified |
| OpenAPI Examples | verified |
| Rate Limit Signal | documented |
| Event Surface Described | derived |
| Agent Skills | derived |
| Well Known Catalog | no |
| Consent Identity | no |
| Agent Card | conformant |
| Dry Run Mode | no |
| Delegated Identity | no |
| Protected Resource Metadata | no |
| Dynamic Client Registration | no |
| Agentic Commerce | no |

## Access

Enterprise — onboarding: unknown, pricing: enterprise, trial: no (confidence: high).

## APIs (3)

- **Refuel Autolabel (Open Source)** — Autolabel is the open-source Python library (pip install refuel-autolabel) to label, clean, and enrich text datasets with any LLM (OpenAI, Anthropic, Google, HuggingFace, vLLM, ...
- **Refuel Cloud API** — The full Refuel Cloud REST API — 108 operations across 77 paths, harvested verbatim from https://cloud-api.refuel.ai/openapi.json on 2026-08-14. Covers projects, datasets and it...
- **Refuel Applications API** — The documented realtime application label surface — the endpoint Refuel's own catalog page publishes as `POST https://cloud-api.refuel.ai/applications/{applicationName}/label`, ...

## MCP servers (1)

- **Refuel.ai** — Refuel serves a live, anonymous, remote MCP endpoint at https://docs.refuel.ai/mcp. It is a DOCUMENTATION server, not a Refuel Cloud API server: the three tools search and read ...

## Agentic access (1)

- **Refuel Ai Agentic Access** — 109 operations · 64 acting

## Security (4)

- **Refuel Ai Authentication** — http · 2 schemes
- **Refuel Ai Domain Security** — TLSv1.3 · HSTS · DMARC
- **Refuel Ai Vulnerability Disclosure** — disclosure policy published
- **Refuel Ai Trust Center** — SOC 2, GDPR

## Plans (1)

- **Refuel Ai Plans Pricing**

## Tags

Artificial Intelligence, LLM, Data Labeling, Data Enrichment, Autolabel, Machine-Learning, Data Quality, Training Data, Fine-Tuning, Data Transformation, Entity Resolution, Content Moderation

---

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