ValGuard.ai logoValGuard.ai

ValGuard + OpenRouter: validate before multi-vendor routing

Use openrouter/ model prefixes through ValGuard Provider Vault, or point a client at ValGuard when OpenRouter is your upstream.

Last verified:

OpenRouter exposes many models behind one API. ValGuard fits in two honest patterns: (1) call ValGuard with openrouter/... model IDs through Provider Vault, or (2) point an OpenAI-compatible client at ValGuard while ValGuard forwards to OpenRouter upstream.

When you need this

You want one validation policy no matter which OpenRouter model answers. Centralize on ValGuard agent slugs instead of per-vendor prompt hacks.

Setup (Provider Vault)

  1. Store your OpenRouter key in Provider Vault.
  2. Create a ValGuard agent with the packs you need.
  3. Call ValGuard with model: "openrouter/anthropic/claude-3.5-sonnet" (example) and X-VG-Agent.
  4. Shadow, tune, enforce.

Code

curl -s https://api.valguard.ai/v1/chat/completions \
  -H "Authorization: Bearer $VG_API_KEY" \
  -H "X-VG-Agent: multi-vendor-gate" \
  -H "Content-Type: application/json" \
  -d '{"model":"openrouter/anthropic/claude-3.5-sonnet","messages":[{"role":"user","content":"Hi"}]}'

Honest limits

  • OpenRouter billing and data handling follow OpenRouter terms when ValGuard forwards there. See subprocessors: LLM providers you configure are not ValGuard subprocessors.
  • ValGuard does not replace OpenRouter model discovery or rate limits.
  • Block and re-ask add latency on top of router and model time.

Related

Next step

Quickstart and supported providers.