GoModel unifies OpenAI requests across providers

This is infrastructure for teams that are already multi-model and want the gateway logic out of application code. The semantic cache is the most interesting part because it goes beyond trivial deduping and can materially cut spend on repetitive prompts.

• –Drop-in OpenAI compatibility lowers migration friction; most apps should only need a base URL change
• –Scoped workflows and aliases make provider routing policy-driven instead of hardcoded per app
• –Semantic caching is the real wedge: similar prompts can reuse answers, not just identical payloads
• –Built-in dashboard, audit logs, and per-user usage tracking make it useful for platform teams, not just hobby projects
• –Best fit is internal AI infrastructure where cost control and observability matter more than a minimal proxy