Interactive mode enables you to review, refine, and re-translate language files on the fly. This powerful feature uses a chat-based workflow to help you achieve the perfect translations.
Enabling Interactive Mode Add the --interactive flag to the translate command:
php artisan translate --interactive
This activates interactive prompts during the translation process.
How It Works
Initial translation
The command translates missing keys as normal
Review prompt
For each language file, you’re prompted to confirm or re-translate
Refinement
If you choose to re-translate, the AI generates new translations with the same context
Iteration
Repeat until you’re satisfied with the results
When to Use Interactive Mode Quality Control Use interactive mode when translation quality is critical:
php artisan translate --name=marketing --language=fr --interactive
This is ideal for:
Marketing copy and user-facing content
Legal or compliance text
Brand-specific terminology
Customer communications
Experimenting with Styles Try different translation approaches:
php artisan translate --name=emails --interactive
You can:
Compare formal vs. informal tones
Test different terminology choices
Refine nuanced translations
Translating Complete Files Interactive mode allows you to re-translate files even when no keys are missing.
When a file has no missing keys, you’ll be prompted:
No missing keys found. Translate the full file again? › Yes / No
Selecting “Yes” will:
Re-translate all keys in the file
Generate fresh translations with full context
Allow you to completely refresh outdated translations
Without interactive mode, files with no missing keys are automatically skipped.
Interactive Workflow Example Here’s what a typical interactive session looks like:
$ php artisan translate --name=auth --language=fr --interactive AI Translator Source: en Languages: fr Domains: auth Translating from en to fr: auth 3 missing keys found: failed, throttle, password ┌ Translating in 1 chunks... (attempt #1 of 3) ────────┐ │ ████████████████████████████████████████ 100% │ └──────────────────────────────────────────────────────┘ Generated translations for auth from en to fr: failed: "Ces identifiants ne correspondent pas à nos enregistrements." throttle: "Trop de tentatives de connexion. Veuillez réessayer dans :seconds secondes." password: "Les mots de passe doivent contenir au moins huit caractères et correspondre à la confirmation." ◆ No missing keys found. Translate the full file again? │ ● Yes / ○ No └
Combining with Other Options Fast Mode + Interactive Use a faster model for quick iterations:
php artisan translate --fast --interactive
This combination is great for rapid prototyping. Use fast mode to quickly generate translations, then review them interactively.
Specific Files + Interactive Focus on particular domains:
php artisan translate --name=auth,passwords --interactive
Multiple Languages + Interactive Review translations for several languages:
php artisan translate --language=fr,es,de --interactive
Interactive mode will prompt for each language-domain combination. With many languages and domains, this can be time-consuming.
Best Practices Start with Non-Interactive For bulk translations:
First run without interactive mode to generate all translations
Then use interactive mode to refine specific files
# Step 1: Bulk translate php artisan translate # Step 2: Refine important files php artisan translate --name=marketing,legal --interactive
Use with Validation Combine validation and interactive mode for a complete workflow:
Check for missing translations
php artisan translate:validate --verbose
Translate interactively
php artisan translate --interactive
Verify results
php artisan translate:validate
Focus on Critical Content Prioritize interactive mode for:
User-facing error messages
Marketing and sales content
Legal and compliance text
Email templates
In-app notifications
Use non-interactive mode for:
Internal error messages
Development/debugging text
Admin panel content
Progress Indicators During translation, you’ll see progress bars showing:
┌ Translating in 5 chunks... (attempt #1 of 3) ────────┐ │ ████████████████░░░░░░░░░░░░░░░░░░░░░░░░ 40% │ └──────────────────────────────────────────────────────┘
Chunks : Large files are split into chunks for processingAttempt : The system makes up to 3 attempts if translation failsProgress : Real-time progress of the current translation
Re-translation Behavior When you choose to re-translate in interactive mode:
The LLM receives the same context as before (full source and target files)
All keys are re-translated, not just missing ones
Previous translations are replaced entirely
The AI may produce different results due to its creative nature
Each re-translation is independent. The AI doesn’t remember your previous interactive session, so results may vary.
Next Steps
Advanced Options Explore power user features and flags
Validation Verify translation completeness