Zero Telemetry
Local VEDA calls never leave your physical workstation. No usage pings, telemetry, or external logging.
Build with Access On AI.
Language models, sovereign local runtimes, and lightweight APIs for intelligent software. Design the path so your first successful API call takes minutes.
Your first successful request
VEDA exposes standard OpenAI-compatible endpoints on localhost:11434 with zero network latency.
import { AccessAI } from '@accessonai/sdk';
// Initialize client (targets sovereign local VEDA by default)
const ai = new AccessAI({
baseURL: 'http://127.0.0.1:11434/v1', // Local offline VEDA runtime
// apiKey: process.env.ACCESS_ON_AI_KEY // Cloud deployment
});
async function main() {
const completion = await ai.chat.completions.create({
model: 'veda-1.0',
messages: [
{ role: 'system', content: 'You are an enterprise sovereign assistant.' },
{ role: 'user', content: 'Extract invoice line items from operational batch #41.' }
],
temperature: 0.1,
});
console.log(completion.choices[0].message.content);
}
main();Zero Cloud Lock-in: The exact same code works locally against VEDA Desktop or against our enterprise cloud endpoints by updating the
baseURL parameter.Standard SDKs
Compatible with official OpenAI clients in TypeScript, Python, Go, and Rust. No proprietary wrappers needed.
Enterprise Support
Contact our engineering team for private VPC deployments, fine-tuning, and high-throughput on-premises clusters.
