Standalone Cloud Proxy Integration Guide

Intutic's proxy engine (intutic-proxy) is a high-performance, OpenAI-compatible proxy gateway written in Rust. It can be deployed in the cloud (such as on GKE) to audit, cache, and govern agent model traffic centrally without installing any workstation shims or developer daemons.

---

🚀 Deployment Topologies

The Rust proxy binary supports three distinct deployment styles:

1. Standalone Cloud Gateway (Zero-Footprint)

• How it works: The proxy runs as a standalone deployment in your cloud cluster (e.g. GKE) and exposes a secure load-balanced endpoint.
• Setup: Point your LLM agent framework or HTTP client directly to the cloud proxy base URL.
• Benefit: Zero footprints on developer machines. All token budget, DLP audits, and caching constraints are applied centrally.

2. Hybrid Gateway Mode

• How it works: Workspace rules files (like .cursorrules or .clinerules) are managed locally via the lightweight workstation sync-daemon to prevent user bypassing, while LLM traffic is routed up to the central GKE proxy deployment.

3. Local Standalone Mode (Air-Gapped)

• How it works: For completely secure, air-gapped, or developer-offline environments, the compiled Rust binary runs locally on the developer's machine as a background service.

---

🛠️ Framework Integration

Connecting any standard LLM framework to the standalone proxy is simple. Just override the base_url parameter and pass your Intutic API key.

1. LangGraph (Python)

In LangGraph, configure the chat model instance to point to the Intutic proxy URL:

from langgraph.graph import StateGraph, START, END
from langchain_openai import ChatOpenAI
from typing import TypedDict

# 1. Initialize LLM pointing to the Intutic proxy gateway
llm = ChatOpenAI(
    model="gpt-4o",
    base_url="https://proxy.intutic.ai/v1",  # GKE Proxy URL
    api_key="your-intutic-api-key"
)

class AgentState(TypedDict):
    input: str
    response: str

def call_model(state: AgentState):
    # This call is governed pre-flight by Intutic
    res = llm.invoke(state["input"])
    return {"response": res.content}

# Compile Graph
builder = StateGraph(AgentState)
builder.add_node("agent", call_model)
builder.add_edge(START, "agent")
builder.add_edge("agent", END)

graph = builder.compile()

2. LangGraph (TypeScript)

Configure the LangGraph state machine runnable context:

import { StateGraph, START, END } from "@langchain/langgraph";
import { ChatOpenAI } from "@langchain/openai";

const model = new ChatOpenAI({
  model: "gpt-4o",
  configuration: {
    baseURL: "https://proxy.intutic.ai/v1", // GKE Proxy URL
    apiKey: "your-intutic-api-key",
  }
});

---

⚡ Interactive Slash Commands

Because the proxy intercepts prompt content pre-flight, developers can invoke interactive slash commands directly inside their chat prompts or agent sessions.

If a prompt begins with /intutic or @intutic, the proxy will process the command immediately and return the output without calling the upstream provider:

# Returns active session budget usage and audit compliance rating
response = llm.invoke("/intutic status")
print(response.content)

Protocol Compliance

To prevent client-side SDK parser failures, the proxy automatically formats the slash command response content to match the requested protocol schema:

• OpenAI SDK: Wraps the payload in a standard chat.completion Choice object.
• Anthropic SDK: Wraps the payload in a standard message JSON block.
• Streaming Content: Streams command outputs using Server-Sent Events (SSE) chunks if stream: true is configured.