Claude Code · Cline · Roo · Kilo
Anthropic base URL + key homelander, or pick provider “OpenAI Compatible”.
BASE_URL https://your-server/openai/v1
API_KEY homelander
MODEL claude-opus-4-8 # or gpt-* (mapped)Anthropic-compatible · runs on a real Claude subscription
A cinematic Claude chat studio and a drop-in proxy — point Claude Code, Cline, OpenCode, or anything OpenAI-compatible at it. Unlimited. Rate-limited. One key: homelander.
01 — What it is
HomeLander AI wraps a genuine Claude Code subscription and exposes it two ways at once.
/v1/messages + OpenAI /v1/chat/completions endpoint that any agent can use as its backend.It forwards to Anthropic with your subscription's OAuth session and Claude Code headers — so traffic rides the plan, not pay-per-token API billing.
02 — Quickstart
Point any Claude Code at the server and run it. The key is always homelander.
export ANTHROPIC_BASE_URL="https://your-server"
export ANTHROPIC_API_KEY="homelander"
claude
Swap https://your-server for this server's URL (your local address or the public tunnel). That's it — you're on the subscription.
03 — Clients
Native Anthropic clients and OpenAI-compatible tools both work. Streaming, tools/function-calling, vision and reasoning_effort are all supported.
Anthropic base URL + key homelander, or pick provider “OpenAI Compatible”.
BASE_URL https://your-server/openai/v1
API_KEY homelander
MODEL claude-opus-4-8 # or gpt-* (mapped)Drop-in for the Messages API.
POST https://your-server/v1/messages
x-api-key: homelanderSet the base URL, use any Claude model id.
base_url = "https://your-server/openai/v1"
api_key = "homelander"04 — Models
Pick per-conversation in the studio, or by id from any client.
claude-fable-5Newest. Vivid, long-form, creative.claude-opus-4-8Deepest reasoning. The heavy hitter.claude-sonnet-4-6Fast, balanced, everyday.claude-haiku-4-5Quick, cheap, snappy.05 — Access
No accounts. No credits. No sign-up. Everyone gets in — capped by a sliding per-IP rate limit (default 60 requests/min) so the subscription stays healthy.
06 — Live telemetry
The studio ships a real-time command-center dashboard: active users, live connections, requests/min, cost projection, cache tokens, latency p50/p95, per-model cost share and throughput charts.
Open the dashboard