llm-trustedrouter

LLM plugin for models available through TrustedRouter, an OpenAI-compatible LLM router.

Installation

llm install llm-trustedrouter

Configuration

Set your API key:

llm keys set trustedrouter
# Paste key here

or set the TRUSTEDROUTER_API_KEY environment variable.

Usage

Models are registered from the public catalog at https://api.trustedrouter.com/v1/models once a key is configured. The list is cached for one hour in trustedrouter_models.json in your LLM user directory.

llm models -q trustedrouter
llm -m trustedrouter/auto "Five names for a pet pelican"

TrustedRouter model IDs are already prefixed (trustedrouter/auto) and are used as-is; other catalog IDs get a trustedrouter/ prefix (trustedrouter/openai/gpt-4o-mini).

Plugin commands:

llm trustedrouter models          # list catalog models
llm trustedrouter models --json   # full catalog JSON
llm trustedrouter refresh         # refresh the cached catalog

Verifying the gateway

The gateway runs as an attested Google Confidential Space workload. The attestation is served at https://api.trustedrouter.com/attestation and the verification procedure is at https://trustedrouter.com/trust (source: https://github.com/Lore-Hex/quill-cloud-proxy).

Development

python -m venv venv && source venv/bin/activate
pip install -e '.[test]'
python -m pytest

Tests mock all HTTP and make no network calls.