✧

Whitepaper v0.1 is live now. Read the Proof of Private, Proof of Green vision.

Grove LogoGrove
IntroductionQuickstartAuthenticationThe Veil, PrivacyThe Watt, GreenGreen CertificateChat completionsPayments (x402)Open Source
Getting Started

Quickstart

Already using the OpenAI SDK? Change one base URL and your app is private + green. Nothing else in your code changes.

OpenAI SDK

Point the client at Grove and pass your key. Every response is computed in a confidential enclave and returns a Green Compute Certificate alongside the answer.

python
from openai import OpenAI

client = OpenAI(
    base_url="https://api.grovecompute.xyz/v1",
    api_key=GROVE_KEY,
)

resp = client.chat.completions.create(
    model="grove-glm",          # or "grove-kimi"
    messages=[{"role": "user", "content": "Hello, Grove"}],
)
print(resp.choices[0].message.content)
print(resp.certificate.serial)  # GCC-...

Over HTTP

Or straight over HTTP with a bearer key, pay-per-call settles over x402:

bash
curl https://api.grovecompute.xyz/v1/chat/completions \
  -H "Authorization: Bearer $GROVE_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "grove-glm",
    "messages": [{"role": "user", "content": "Hello"}],
    "stream": false
  }'

Embeddings

Create private embeddings for retrieval over your own data, same drop-in interface:

python
client.embeddings.create(
    model="grove-embed",
    input=["confidential document chunk"],
)

Next steps

  • Authentication, wallet sign-in for the dApp, bearer keys for the API.
  • The Veil / The Watt, how privacy and carbon are enforced per job.
  • Chat completions, full request/response, streaming, and models.
Model names are stable aliases (grove-glm, grove-kimi, grove-embed), the network picks the cheapest, greenest node that meets your quality bar.
← IntroductionAuthentication →