Skip to content
Hugging Face → EvalGuard

Load your eval datasets. Run the full platform on them. 

Hugging Face is where your benchmark and eval-result datasets live. EvalGuard imports those rows as spans so you can analyze, re-score (200+ scorers), red-team, and secure them — all in one workspace.

SOC 2 evidence engineISO 42001 mappedEU AI ActGDPR

A dataset was the start

What you keep — and what you gain

Keep
  • Dataset rows as records (Datasets-Server JSON or a JSONL split)
  • Input — question / prompt per row
  • Output — answer / prediction per row
  • Expected / reference answer
  • Model column + metric score
  • Every other column preserved, nothing lost
Gain
  • + Real-time LLM firewall (input + output)
  • + 300+ red-team plugins with threat-feed sync
  • + 200+ eval scorers (LLM-as-judge, pairwise, rubric)
  • + Prompt IDE + optimizer across 90+ providers
  • + Managed BYOK gateway with semantic caching
  • + 50 compliance frameworks + tamper-evident audit log

Bring your data

Import a Hugging Face eval dataset

Point the CLI at a Datasets-Server slice (the /rowsJSON) or a JSONL / JSON split you've downloaded. The mapping is deliberately conservative — common eval fields become spans and every other column is preserved — and re-running the same import never double-counts, because spans use a stable content hash.

# Convert a Hugging Face eval dataset to neutral-shape spans:
npx @evalguard/cli import:traces --from huggingface hf-dataset.json --output spans.json

Prints an import summary (spans imported / duplicates skipped / parse errors) and writes the neutral spans to spans.json. This is a dataset import for eval / benchmark result rows — not a live trace feed.

What importsWhere it lands
Dataset rows (question / answer / etc.)One EvalGuard span each
Input (input / question / prompt)span.input
Output (output / answer / prediction)span.output
Expected / referencespan.expected
Modelspan.model
Metric scorespan.score
Latency + cost (when present)span.durationMs / span.costUsd
Every other columnspan.attributes (huggingface.row.* namespace)

Hugging Face eval datasets don't carry token usage, so prompt / completion token counts are left blank on import. A row with passed: false imports with status: error so your failing cases stay visible.

One platform, one bill

Load a dataset. Score, secure, optimize — everywhere.

A dataset is the hook. The platform is why you stay.

Start free