How to Translate Content with AI in conbase.ai

Published on July 08, 2025

By Daniel Manco

Introduction

  • The Goal:
    • By the end of this guide, you'll be able to automatically create accurate translations for hundreds or thousands of content records in minutes, transforming your localization workflow.
  • The Problem:
    • Manually translating each piece of content is slow, expensive, and nearly impossible to scale. Teams often copy and paste text into free tools, then spend hours fixing mistakes, blocking product launches and global campaigns.
  • The Solution:
    • Conbase.ai's AI-powered pipeline approach lets you translate at scale using custom prompts that understand your brand voice and terminology. You create tailored workflows that match your exact requirements.
  • Prerequisites:
    • A Conbase.ai account
    • A CSV file containing your content with columns like id, source_text, target_language
    • Glossary or style guide (optional but recommended)

Are you here for the first time ? Then watch a full conbase.ai workflow in action

This video walks you through a complete conbase.ai workflow, from input to final output, so you can see exactly how the platform works and how to get the most out of its features.

Step 1: Set Up Your Project and Import Data

Create a New Project

Log in and click New Project. Name it something like Content Translation - March 2025.

Upload Your Content CSV

Drag your CSV into the uploader. Conbase.ai detects headers automatically.

  • Use clear column names such as source_text and target_language
  • Remove stray HTML tags or broken encoding
  • Add a column for brand_terms if you have a glossary

Step 2: Build Your AI Translation Pipeline

Add an Action Step

Click Add Step, then Create New. Name the step Translate.

Set Up Your Custom Prompt: A 3-Part Recipe

Part 1: Define the Instruction

You are a professional translator. Translate the provided source text into the requested target language.
Style: keep the original tone and follow the brand glossary if supplied.
Do not add extra notes.
Return only the translated text.

Part 2: Select the Context

Source Text: {source_text}
Target Language: {target_language}
Brand Terms: {brand_terms}

Part 3: Define the Output

  • Create a new column called translated_text
  • Add two output examples:

Output Column: translated_text
Output Example: 
If spanish: Hola, bienvenido a nuestra aplicación.
If german: Willkommen in unserer App.

Click Save Changes.

Step 3: Configure and Run the Pipeline

Fine-Tune AI Settings

  • Model: "GPT-4o"
  • Processing Mode: "Instant" for testing and for fast execution, "Scheduled" for large batches and 50% less API costs
  • Integration: "Fill if empty" leaves available data untouched

Select Rows and Execute

  1. Select 10 rows for a test run
  2. Review results in the cell viewer
  3. Tweak the prompt if you spot errors
  4. When happy, select all rows and run the full pipeline

Step 4: Review and Export Your Results

Preview the Translations

  • Spot-check random rows for tone consistency
  • Verify that glossary terms are honored
  • Fix any outliers by rerunning single rows

Export Your CSV

  • Click Export in the bottom right corner
  • Download the CSV with the new translated_text column
  • Import into your CMS or send to your localization team for final QA

Conclusion and Next Steps

You just automated translation for your entire content library. Faster launches and consistent voice are now within reach.

  • Track time saved by comparing manual and automated efforts
  • Measure quality using native speaker reviews
  • Reuse this pipeline whenever new content arrives

Book Your Free Conbase.ai Demo and see how else you can streamline content operations.

Ⓒ Copyright. All rights reserved by conbase.ai