Aiursoft Ollama Gateway

Aiursoft Ollama Gateway gives teams controlled access to shared local LLM servers. It connects AI applications and coding assistants to Ollama or OpenAI-compatible backends through a central gateway.

• Client compatibility: Accept requests using Ollama, OpenAI Chat Completions, OpenAI Responses or Anthropic Messages APIs, with compatibility for Claude Code, Open-WebUI, Opencode, Roocode and Qwen Code.
• Shared access controls: Manage separate API keys, permissions, team roles and rate limits for users and applications.
• Virtual models and routing: Give models stable aliases with persistent system prompts and enforced generation settings, with multiple backends, load balancing and fallback.
• Auditing and monitoring: Store request and response audits in ClickHouse and view traffic, latency and usage through the supplied Grafana dashboard template.
• Embedding models: Manage embedding models for document retrieval applications, alongside chat models; a keep-alive feature keeps underlying Ollama models loaded between requests.

The gateway is open source under the MIT license and runs on an Ubuntu server or in Docker, with a browser dashboard. Inference takes place on the chosen backend. Local Ollama uses your hardware, while requests routed to an external provider go to that service.

Protocol translation has limits: matching protocols retain unknown fields and streaming events, while translated routes preserve supported content. Stateful OpenAI Responses features require a Responses backend. Anthropic Messages is a client-facing API; the gateway has no physical Anthropic backend adapter.