GoModel

GoModel is a self-hosted AI gateway for developers and platform teams sharing access to local models and cloud providers. It centralizes provider choice and usage policies behind APIs compatible with OpenAI and Anthropic SDKs.

• Model routing: Connect Ollama, vLLM or SGLang alongside OpenAI, Anthropic and Gemini, with stable model aliases, load balancing and automatic failover to backup providers.
• Caching and conversation continuity: Reuse exact or semantically similar responses and keep a conversation assigned to the same provider target and key.
• Access and spending controls: Issue managed API keys, restrict model access, enforce budgets and rate limits, and apply different caching, logging and guardrail policies to different workloads.
• Usage and audit records: Attribute usage to teams, tenants or features through the built-in dashboard, with request logs, provider status, Prometheus metrics and OpenTelemetry traces.
• Beyond text chat: Route embeddings, speech generation, transcription and image requests through the gateway, or bring MCP servers together behind one authenticated endpoint.

GoModel runs on macOS, Windows and Linux, including Docker deployments, and includes its admin dashboard in the same binary. The open-source core uses the MIT license; a separate commercial distribution adds licensed extensions. The gateway runs on your hardware, but requests sent to cloud providers go to those services for processing. Semantic caching requires a supported vector store such as Qdrant, pgvector, Pinecone or Weaviate.