TrustedRouter / Renewable-powered inference
Green tokens.
Your next million tokens.
Powered by renewable energy.
Choose AI inference powered by 100% renewable electricity.
Same API. Same workflow. A different source of power.
One model name
Make energy part
of the request.
Set your model to trustedrouter/green. TrustedRouter selects an eligible model and keeps every fallback inside the renewable-powered provider pool.
import os
from openai import OpenAI
client = OpenAI(
base_url="https://api.trustedrouter.com/v1",
api_key=os.environ["TRUSTEDROUTER_API_KEY"],
)
for chunk in client.chat.completions.create(
model="trustedrouter/green",
messages=[{"role": "user", "content": "What should we build next?"}],
max_tokens=512,
stream=True,
):
if chunk.choices:
print(chunk.choices[0].delta.content or "", end="", flush=True)One model name, a growing renewable-powered provider pool. Provider routing docs.
Ready to route
Choose your model.
Keep your energy choice.
The current eligible catalog. Model availability and prices refresh directly from providers. More green providers are coming.
Follow the evidence
A power choice.
A clear boundary.
- Inference electricity*
- Green routes require a published commitment to 100% renewable inference electricity. The current provider's energy policy was reviewed September 14, 2026. This provider-declared claim covers inference electricity, rather than model training, hardware manufacturing, your network or TrustedRouter's routing infrastructure.
- Provider privacy
- The current provider's zero-data-retention policy says prompts and outputs are discarded after processing. Privacy and renewable energy are separate commitments. Each provider's policies remain visible in the catalog as the pool grows.
- Gateway proof
- TrustedRouter handles routing inside an attested gateway. Your agent can verify the running gateway. That proof covers gateway code and hardware, rather than the electricity source or the provider's inference hardware.
- Routing guarantee
- The green alias restricts eligible providers. When none can serve your request, the request fails closed. Energy requirements remain in place through model fallbacks, provider preferences and usage settlement.
A small change
in where your AI runs.
Questions
What does 100% renewable mean here?
Eligible providers publish a commitment to 100% renewable inference electricity. This provider-declared claim covers electricity used by inference servers. Model training, hardware manufacturing, your network and routing infrastructure have separate energy footprints.
What happens if a green route is unavailable?
TrustedRouter can try another eligible model route within the renewable-powered provider pool. If none can serve the request, it returns an error. It does not silently switch to an unqualified provider.
Does green mean end-to-end confidential inference?
Energy sourcing and hardware confidentiality are different. Green tokens select for renewable inference electricity. Check each provider's privacy and hardware-verification status in the catalog; the green alias itself makes an energy commitment, rather than a confidential-compute commitment.
How is usage priced?
The selected model's provider rate and TrustedRouter's normal inference fee apply. Published prices are in US dollars, with currency conversion where needed. Current model prices are in the catalog. Green routing is included in the normal fee.