Skip to content

Connect your agents to confidential hardware

Use OpenClaw, CrewAI, LangChain, or any OpenAI SDK. One line change: every LLM call goes to a model inside Intel TDX at our inference provider. Requests pass through our relay, which sees them in clear (TLS ends at our host). To keep plaintext off our servers, run your own inference server in a Confidential VM and terminate TLS inside its trust domain.

1
Create an account
Sign up at voltagegpu.com
→
2
Set the base URL
api.voltagegpu.com/v1/confidential
→
3
Run your agent
Every LLM call now goes to a model running in Intel TDX. Zero code rewrite.

OpenClaw

Open source CLI

Run any OpenClaw agent on confidential hardware. One config change.

Best for: CLI-first agent workflows, local development, rapid prototyping

# ~/.openclaw/config.yaml
providers:
  - name: voltagegpu-confidential
    type: openai
    base_url: https://api.voltagegpu.com/v1/confidential
    api_key: vgpu_YOUR_API_KEY
    model: contract-analyst  # or any of the 9 agents

# That's it. Every LLM call now runs in a TDX enclave.

CrewAI

Python framework

Multi-agent orchestration with CrewAI, now running inside Intel TDX enclaves.

Best for: Multi-agent teams, role-based workflows, complex research tasks

from crewai import Agent, Task, Crew
from langchain_openai import ChatOpenAI

# One line change: point to VoltageGPU TDX.
# Requests pass through our relay, which sees them in clear (TLS ends at our host).
# To keep plaintext off our servers, run your own inference server in a
# Confidential VM and terminate TLS inside its trust domain.
llm = ChatOpenAI(
    base_url="https://api.voltagegpu.com/v1/confidential",
    api_key="vgpu_YOUR_API_KEY",
    model="financial-analyst",  # or any of the 9 agents
)

analyst = Agent(role="Senior Financial Analyst", llm=llm, ...)
crew = Crew(agents=[analyst], tasks=[...])
crew.kickoff()  # The model runs inside Intel TDX at our inference provider

LangChain / LangGraph

Python / JS framework

LangChain and LangGraph chains with confidential inference. Drop-in replacement. Requests pass through our relay, which sees them in clear (TLS ends at our host). To keep plaintext off our servers, run your own inference server in a Confidential VM and terminate TLS inside its trust domain.

Best for: RAG pipelines, retrieval QA, document processing, complex chains

from langchain_openai import ChatOpenAI

# Drop-in replacement, same OpenAI SDK, model inside Intel TDX.
# Requests pass through our relay, which sees them in clear (TLS ends at our host).
# To keep plaintext off our servers, run your own inference server in a
# Confidential VM and terminate TLS inside its trust domain.
llm = ChatOpenAI(
    base_url="https://api.voltagegpu.com/v1/confidential",
    api_key="vgpu_YOUR_API_KEY",
    model="compliance-officer",
)

# Use in any chain, agent, or RAG pipeline
from langchain.chains import RetrievalQA
qa = RetrievalQA.from_chain_type(llm=llm, retriever=retriever)
result = qa.invoke("Does this policy comply with GDPR Art. 25?")

Any OpenAI SDK

Python / Node.js / any language

Any OpenAI-compatible SDK works: change the base URL. Requests pass through our relay, which sees them in clear (TLS ends at our host). To keep plaintext off our servers, run your own inference server in a Confidential VM and terminate TLS inside its trust domain.

Best for: Custom apps, existing codebases, any language with OpenAI SDK support

# Python
from openai import OpenAI
client = OpenAI(
    base_url="https://api.voltagegpu.com/v1/confidential",
    api_key="vgpu_YOUR_API_KEY",
)
response = client.chat.completions.create(
    model="contract-analyst",
    messages=[{"role": "user", "content": "Review this NDA..."}],
)

# Node.js, identical pattern
# import OpenAI from 'openai';
# const client = new OpenAI({ baseURL: "https://api.voltagegpu.com/v1/confidential", apiKey: "vgpu_..." });

Ready to connect your agent?

Browse pre-built templates or check the API docs to integrate your own.