Most LLM gateways assume you're paying per-token via API key. SheLLM's primary use case is the opposite: you already have CLI subscriptions (Claude Max, Gemini AI Plus, OpenAI Enterprise) and want to expose them as a regular HTTP API to your apps — without burning API quota.
Local execution bridge for Anthropic's power tools.
Seamless integration with Google's command line suite.
OpenAI's enterprise code generation command line.
Extreme throughput for large-scale inference tasks.
Start your gateway in seconds. SheLLM handles the routing, you handle the queries.
# Clone and install
$git clone https://github.com/rodacato/SheLLM.git && cd SheLLM
$bash scripts/setup/dev.sh
# Verify environment
$npm run check:env
# Start the gateway
$shellm start
# Send a request
$curl http://localhost:6100/v1/chat/completions \
-d '{"model":"claude","messages":[{"role":"user","content":"Hello"}]}'
| Feature | SheLLM | LiteLLM | OpenRouter | Portkey |
|---|---|---|---|---|
| CLI subscriptions as backends | ✓ | ✗ | ✗ | ✗ |
| OpenAI-compatible endpoint | ✓ | ✓ | ✓ | ✓ |
| Anthropic-compatible endpoint | ✓ | ✗ | ✗ | ✗ |
| Built-in admin dashboard | ✓ | ✗ | ✗ | ✓ |
| Self-hosted, SQLite only | ✓ | ✗ | ✗ | ✗ |