I want to…
Connect a local runtime
Expose what you already run — without opening a port.
OllamaLM Studiollama.cppvLLM
→Use Tokios with an agent or app
Point a coding agent or chat app at your local model.
Claude CodeCodexCursor
→Choose a model
Find what fits your hardware, and which family fits your task.
Fit finderBy task
→Get started in 3 steps
01
Run a model
Any backend that speaks Chat Completions, on your own hardware.
02
Connect Tokios
Install the connector and register your model in under 5 minutes.
03
Use your endpoint
Point any OpenAI- or Anthropic-compatible client at your endpoint.
Verified compatibility
Check what works out of the box — View the complete matrix →Client
Ollama
vLLM
LM Studio
Claude Code
Verified
Verified
Verified
OpenAI Codex
Verified
Verified
Verified
OpenAI SDK
Verified
Verified
Verified
Anthropic SDK
Verified
Verified
Verified