{"slug":"set-up-llm-clients","title":"Set up LLM clients","tags":["tailscale"],"agent_summary":"Last validated: Apr 9, 2026","trigger_phrases":[],"runnable":false,"markdown":"\r\n# Set up LLM clients\r\n\r\nLast validated: Apr 9, 2026\r\n\r\nAperture by Tailscaleis currently [in alpha](https://tailscale.com/docs/reference/tailscale-release-stages#alpha).\r\n\r\nAfter [setting up Aperture](https://tailscale.com/docs/aperture/get-started) and [configuring your providers](https://tailscale.com/docs/aperture/set-up-providers), set up your LLM clients to route requests through your Aperture instance.\r\n\r\nAperture sits between your LLM clients and your upstream providers and routes requests from each client to the appropriate provider. When a client sends a request specifying a model name, Aperture routes it to the provider that serves that model and automatically injects authentication. Any LLM client that supports a custom base URL can use Aperture.\r\n\r\nUse the following guides to configure an LLM client to use your Aperture instance. Refer to the [provider compatibility reference](https://tailscale.com/docs/aperture/provider-compatibility) for the full list of supported providers and API formats.\r\n\r\n[**Use Claude Code with Aperture** \\\\\r\n\\\\\r\nConfigure Claude Code to route requests through your Aperture proxy.](https://tailscale.com/docs/aperture/how-to/use-claude-code) [**Use Codex with Aperture** \\\\\r\n\\\\\r\nConfigure OpenAI Codex to route requests through your Aperture proxy.](https://tailscale.com/docs/aperture/how-to/use-codex) [**Use OpenCode with Aperture** \\\\\r\n\\\\\r\nConfigure OpenCode to route requests through your Aperture proxy.](https://tailscale.com/docs/aperture/how-to/use-opencode) [**Use OpenAI-compatible tools with Aperture** \\\\\r\n\\\\\r\nConfigure Gemini CLI, Roo Code, Cline, and other OpenAI-compatible tools to route requests through Aperture.](https://tailscale.com/docs/aperture/how-to/use-openai-compatible-tools)\r\n\r\n![Project Logo](https://cdn.brandfetch.io/tailscale.com/fallback/lettermark/theme/dark/h/256/w/256/icon?c=1bfwsmEH20zzEfSNTed)\r\n\r\nAsk AI\r\n\r\nreCAPTCHA\r\n\r\nRecaptcha requires verification.\r\n\r\nprotected by **reCAPTCHA**\r\n","html":"

Set up LLM clients

\n

Last validated: Apr 9, 2026

\n

Aperture by Tailscaleis currently in alpha.

\n

After setting up Aperture and configuring your providers, set up your LLM clients to route requests through your Aperture instance.

\n

Aperture sits between your LLM clients and your upstream providers and routes requests from each client to the appropriate provider. When a client sends a request specifying a model name, Aperture routes it to the provider that serves that model and automatically injects authentication. Any LLM client that supports a custom base URL can use Aperture.

\n

Use the following guides to configure an LLM client to use your Aperture instance. Refer to the provider compatibility reference for the full list of supported providers and API formats.

\n

Use Claude Code with Aperture \\\r\n\\\r\nConfigure Claude Code to route requests through your Aperture proxy. Use Codex with Aperture \\\r\n\\\r\nConfigure OpenAI Codex to route requests through your Aperture proxy. Use OpenCode with Aperture \\\r\n\\\r\nConfigure OpenCode to route requests through your Aperture proxy. Use OpenAI-compatible tools with Aperture \\\r\n\\\r\nConfigure Gemini CLI, Roo Code, Cline, and other OpenAI-compatible tools to route requests through Aperture.

\n

\"Project

\n

Ask AI

\n

reCAPTCHA

\n

Recaptcha requires verification.

\n

protected by reCAPTCHA

\n"}