Multi-route high-availability

Key features

• Model groups — group an ordered list of provider routes that serve the same model under one virtual provider name, with per-route quarantine tuning.
• Keep your agent working when one provider route stumbles.
• It does not bypass quotas, rate limits, or fair-use caps.
• The host never reads or writes key material
• Round-robin allocation — new requests spread across the healthy routes of the active group