Move your inference logs. Keep your eval scores.
Athina monitors and evaluates your LLM inferences. EvalGuard imports those logs — eval scores and all — and adds an LLM firewall, red team (300+ plugins), and a BYOK gateway on the same traffic, in one workspace.
Eval + monitoring was the start
What you keep — and what you gain
- ✓ Inference logging (prompt / response capture)
- ✓ Model + provider (inferred from the model when absent)
- ✓ Token usage — prompt / completion / total
- ✓ Cost + latency per inference
- ✓ Per-inference eval scores (groundedness, etc.)
- ✓ Environment + session grouping
- + 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
Backfill your history
Bring your Athina logs along
Export your inference logs from Athina (the dashboard JSON download, or the logging API), then convert them with the EvalGuard CLI. No record is dropped — anything the importer can't parse is surfaced as an error — and re-running the same import never double-counts, because spans use a stable content hash.
Prints an import summary (spans imported / duplicates skipped / parse errors) and writes the neutral spans to spans.json. This is a trace / observability import — it moves your inference history and its recorded scores, not your Athina prompt or evaluator configuration.
| What imports | Where it lands |
|---|---|
| Inference logs | One EvalGuard span each, grouped by session id |
| Model + provider | span.model / span.provider (provider inferred when absent) |
| Prompt / response | span.input / span.output |
| Token usage (prompt / completion / total) | span.promptTokens / completionTokens / totalTokens |
| Cost | span.costUsd |
| Latency | span.durationMs (response_time) |
| Athina eval scores + RAG context + metadata | span.attributes (athina.* namespace) |
Athina's recorded eval results (name, score, pass / fail) are preserved under athina.eval.* so you can chart your old scores next to fresh EvalGuard runs.
One platform, one bill
Evaluate once. Secure, monitor, optimize — everywhere.
Eval + monitoring is the hook. The platform is why you stay.
Start free