Helicone joined Mintify — if your priorities just changed, read on

Helicone did logging. EvalGuard does logging + five more things.

Drop-in proxy replacement — change the base URL, keep your SDK code. You get request/response logging, cost tracking, and caching just like Helicone. Plus red-team security (249 plugins), LLM firewall, evals (106 scorers), and prompt IDE in the same workspace.

What you keep — and what you gain

Keep
  • ✓ Drop-in proxy — change one URL, keep your SDK
  • ✓ Full request / response logging
  • ✓ Cost tracking per model, per user
  • ✓ Response caching (semantic + exact)
  • ✓ Custom tags / properties for segmentation
  • ✓ Rate limiting per key
  • ✓ Generous free tier (10K traces/month)
Gain
  • + Real-time LLM firewall (input + output)
  • + 249 red-team plugins with threat-feed sync
  • + 106 eval scorers (LLM-as-judge, pairwise, rubric)
  • + Prompt IDE + optimizer across 87 providers
  • + OTel trace correlation web→worker→LLM
  • + Drift detection + auto-reeval on regression
  • + 33 compliance frameworks + audit log

Proxy URL mapping

Replace the base URL and change one header. Your SDK code, your provider keys, and your custom metadata stay exactly the same.

HeliconeEvalGuardNotes
https://oai.helicone.ai/v1https://evalguard.ai/api/v1/gateway/proxySame drop-in base-URL pattern; EvalGuard routes to the right provider based on the model name.
https://anthropic.helicone.aihttps://evalguard.ai/api/v1/gateway/proxySame endpoint — claude-* model names auto-route to Anthropic.
Authorization: Bearer $OPENAI_KEYAuthorization: Bearer $PROVIDER_KEYOr let EvalGuard hold your provider keys and drop the header entirely.
Helicone-Auth: Bearer $HELICONE_KEYX-EvalGuard-Key: eg_...Proxy auth header — get yours at /dashboard/settings.
Helicone-Property-*: metadataX-EvalGuard-Tags: tag1,tag2Custom tags work identically.

One-line migration

// Before — Helicone
- baseURL: "https://oai.helicone.ai/v1",
- defaultHeaders: { "Helicone-Auth": `Bearer ${HELICONE_KEY}` },
// After — EvalGuard
+ baseURL: "https://evalguard.ai/api/v1/gateway/proxy",
+ defaultHeaders: { "X-EvalGuard-Key": EG_KEY },

Log once. Secure, eval, optimize — everywhere.

Observability is the hook. The platform is why you stay.

Start free — 10K traces / month
Migrate from Helicone — observability + everything else | EvalGuard