Developer Reference v2.4.0
The Neural SDK.
Connect your legacy enterprise systems to high-performance private LLMs. Our SDK handles the complex orchestration of vector retrieval, prompt safety, and token efficiency automatically.
Distributed Inference Architecture
Our global edge network reduces LLM latency by processing non-sensitive tokens at the nearest node.
01. Integration
Install the core library using your preferred package manager:
$ npm install @gurugram-ai/core-sdk02. Secure Inference
Inference_Request.ts
const response = await ai.chat.completions.create({
model: "gurugram-ultra-v1",
messages: [{ role: "user", content: "Analyze market debt." }],
// Enterprise Security Layer
security: {
pii_redaction: true,
adversarial_check: true
},
// Latency Optimization
stream: true
});Sub-100ms TTFT
Optimized speculative decoding for near-instant first token generation.
SOC2 Data Guard
Automated compliance mapping for every inference call made to the model.