# Probabl

**Canonical:** https://apis.io/providers/probabl/  
**Website:** https://probabl.ai/  
**APIs profiled:** 1

Probabl (styled ":probabl.") is the French, Inria-spun-out company behind scikit-learn — "the Tabular AI company by the creators of scikit-learn" — founded in 2023 to develop, maintain and commercially sustain the open-source Python data-science stack (scikit-learn, skrub, skore). Its commercial product is Skore: an open-source Python library (`skore`) that turns model training into structured, auditable artifacts — estimator reports, cross-validation reports, comparison reports, data and metric diagnostics — plus Skore Hub, a hosted collaboration platform where teams push, compare and track those artifacts across workspaces and projects. Skore Hub exposes a public REST API (Skore Hub, OpenAPI 3.1.0, 67 operations) covering identity/OAuth, workspaces, members and API keys, projects, estimator/cross-validation/comparison reports, artifacts, project goals, and OpenAI/Anthropic-compatible agent endpoints. Probabl also publishes a first-party Agent Skills package (probabl-skills, BSD-3-Clause) of 14 data-science skills for coding agents such as Claude Code and Cursor, a `skore` CLI, training via Skolar, and the Scikit-learn Central ecosystem explorer.

## Kin Score — 39.7 / 100 (developing)

Scored 2026-09-08 under rubric 0.20.0. Trend: flat (+0.0 from 39.7).

| Facet | Score |
|---|---|
| Discoverability | 68.5 |
| Contract Quality | 46.4 |
| Governance | 0.0 |
| Contract Governance | 0.0 |
| Operational Transparency | 26.3 |
| Developer Ergonomics | 71.4 |
| Commercial Clarity | 21.1 |
| Access Clarity | 21.1 |

## Agent readiness — 36.9 (agent-ready)

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

## Access

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

## APIs (8)

- **Probabl Anthropic Compatible Agent API** — The Anthropic Compatible Agent API from Probabl — 1 operation(s) for anthropic compatible agent.
- **Probabl Health API** — The Health API from Probabl — 1 operation(s) for health.
- **Probabl Identity API** — The identity API from Probabl — 22 operation(s) for identity.
- **Probabl Liveness API** — The Liveness API from Probabl — 1 operation(s) for liveness.
- **Probabl LLM Provider API** — The LLM Provider API from Probabl — 3 operation(s) for llm provider.
- **Probabl OpenAI Compatible Agent API** — The OpenAI Compatible Agent API from Probabl — 3 operation(s) for openai compatible agent.
- **Probabl Projects API** — The projects API from Probabl — 19 operation(s) for projects.
- **Probabl Readiness API** — The Readiness API from Probabl — 1 operation(s) for readiness.

## MCP servers (1)

- **Probabl MCP Server**

## Security (1)

- **Probabl Domain Security** — TLSv1.3 · DMARC

## Tags

Company, Open-Source, Machine-Learning, Data Science, scikit-learn, MLOps, Model Evaluation, Experiment Tracking, Agent Skills, Artificial Intelligence, Python, France

---

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