Hugging Face Hub / Inference
- Fair 2.7
- Easy 3.7
- Secure 2.3
FES Scores 9
| Pillar | Criterion | Score | Evidence |
|---|---|---|---|
| Easy | Day-one UX Can non-technical staff use it without a project? | 3 | Typical onboarding for this product category. Source |
| Easy | Operate Admin burden, updates, backups and language coverage. | 3 | Vendor operates the hosted service. Source |
| Easy | Integrate SSO, mail/CalDAV, APIs and common NGO connectors. | 5 | Integrations vary; check current APIs and connectors. Source |
| Fair | Lock-in / exit Can you leave with your data in open formats, at known cost? | 2 | Rule baseline (rubric_v1): global SaaS APIs/export exist, but suite gravity and remapping remain. Source |
| Fair | Sovereignty EU legal entity, EU hosting, and exposure to extra-EU lawful access? | 1 | Rule baseline (rubric_v1): non-EU legal entity + global SaaS — high extra-EU lawful-access exposure. Source |
| Fair | Openness Open source or open standards? Are independent implementations possible? | 5 | Rule baseline (rubric_v1): license classified as open-source family. Source |
| Secure | IT security Authentication, encryption, patch cadence and independent audit. | 3 | Hugging Face offers a Data Processing Addendum (DPA) outlining security commitments. Source |
| Secure | User security 2FA, roles, phishing resistance and safe defaults. | 2 | Privacy policy mentions user data protection but lacks detailed security controls. Source |
| Secure | Personal data GDPR fit: DPA, residency, sub-processors and Schrems II posture. | 2 | Privacy policy describes data collection and processing but does not specify encryption standards. Source |
Indicative scores from public information — how we score.