Getting Started

Get VersionR integrated into your app in under 5 minutes.

Create a project

Log in to app.versionr.ai and create a new project. Give it a name and slug — the slug is how the SDK identifies your project.

Create your first prompt

Inside your project, click New Prompt. Give it:

A name (e.g. Interviewer System)
A slug (e.g. interviewer-system) — this is what you’ll reference in your code

Write your prompt content. Use {{double_curly_braces}} for any dynamic values:

You are an experienced technical interviewer.
Your candidate is {{candidate_name}}, applying for a {{job_role}} role.
Difficulty level: {{difficulty}}.

Conduct a professional but friendly interview.

Click Save Version — this creates v1 of your prompt.

Deploy to an environment

On the Versions tab, click Deploy → development on v1. Your prompt is now live in the development environment.

Get your API key

In your project, go to the API Keys tab and create a key. Copy it — you’ll only see it once.

Install the SDK

npm install versionr
# or
pnpm add versionr
# or
yarn add versionr

Fetch your prompt

import { VersionR } from 'versionr'
 
const pd = new VersionR({
  apiKey: process.env.VERSIONR_API_KEY,
})
 
// Fetch the active prompt for the development environment
const prompt = await pd.get('interviewer-system', { env: 'development' })
 
console.log(prompt.content)
// → "You are an experienced technical interviewer..."

Render with variables

const rendered = await pd.render('interviewer-system', {
  env: 'development',
  variables: {
    candidate_name: 'Eoin',
    job_role: 'Senior Engineer',
    difficulty: 'hard',
  },
})
 
console.log(rendered.content)
// → "You are an experienced technical interviewer.
//    Your candidate is Eoin, applying for a Senior Engineer role.
//    Difficulty level: hard...."

Pass it to your LLM

import OpenAI from 'openai'
 
const openai = new OpenAI()
 
const response = await openai.chat.completions.create({
  model: rendered.model,           // model is versioned alongside your prompt
  temperature: rendered.temperature,
  messages: [
    { role: 'system', content: rendered.content },
    { role: 'user', content: userMessage },
  ],
})

What’s next?

Deploy to production — promote a version to production
Log results — send results back to VersionR for visibility
A/B testing — split traffic between two prompt variants
SDK Reference — full API reference

SDK reliability — if VersionR is unreachable, the SDK returns a cached version of your prompt and never crashes your app. See error handling.

Introduction Projects