Web Google Translate Guide

🌐 Using Web Google Translate with RenLocalizer

Date: February 8, 2026 | Version: 2.6.7+
Author's Note: This guide is prepared specifically for users utilizing the free web-based Google Translate service (translate.google.com).

---

📋 Table of Contents

1. What is Web Google Translate?
2. Technical Limitations
3. RenLocalizer Solutions
4. Configuration
5. Best Practices
6. Troubleshooting
7. Improvement Roadmap

---

🌐 What is Web Google Translate?

Definition

Web Google Translate = The free, browser-based translation service accessed via translate.google.com.

Comparison

Feature	Web Version (Free)	Paid Cloud API v2
Cost	✅ Free	❌ Paid (~$20 per 1M chars)
HTML Protection	❌ None	✅ format=html
translate="no" Support	❌ No support	✅ Full support
Speed	⚠️ Slower (Rate limiting)	✅ Very Fast
Reliability	⚠️ Variable	✅ Guaranteed SLA

---

⚠️ Technical Limitations

Problem #1: Lack of HTML Protection

Sent:      "Hello [player_name] {color=#fff}text{/color}"
Google's Perception: Everything is plain text to be translated.
↓
Result:    "[player OUTSIDE_NAME {color = #fff} text {/ color}" ← CORRUPTED!

Reason: The web-based Google Translate:

• Does not understand the format=html parameter.
• Ignores HTML attributes like translate="no".
• Focuses strictly on plain text translation.

Problem #2: Spaced Token Corruption

Google Translate often adds arbitrary spaces to technical tokens during translation:

Sent by RenLocalizer:   VAR0
Google translation:     VAR 0  ← Space added!
Result:                 Cannot restore VAR 0 to the original variable.

Problem #3: Strict Rate Limiting

• Web-based = Rate limits applied per IP.
• Stricter restrictions when using standard VPNs/Proxies.
• Risk of temporary bans: 429 Too Many Requests.

---

✅ RenLocalizer Solutions

Solution #1: Token-Based Protection (Default for Web)

Principle: Use tokens instead of HTML tags during the translation phase.

# RenLocalizer sends protected text:
"Hello VAR0 TAG0textTAG1"

# Placeholder dictionary (stored locally):
{
    'VAR0': '[player_name]',
    'TAG0': '{color=#fff}',
    'TAG1': '{/color}'
}

# Google translates:
"Hello VAR0 TAG0textTAG1"

# RenLocalizer restores:
"Hello [player_name] {color=#fff}text{/color}"

Setting:

{
    "use_html_protection": false
}

Solution #2: Spaced Token Recovery (NEW v2.6.7+)

Problem: Google changes VAR0 → VAR 0.

Solution: Phase 0.5 Pre-processing logic.

# Inside restore_renpy_syntax():
spaced_pattern = re.compile(r'(VAR|TAG|ESC_OPEN|ESC_CLOSE|XRPYX[A-Z]*)\s+(\d+|[A-Z_]*)')
# Converts "VAR 0" back to "VAR0" before restoring the original syntax.

Test Results:

"VAR 0 created" → "[player_name] created" ✅ FIXED
"TAG 5 text" → "{b}text" ✅ FIXED
"ESC_OPEN 2 code" → "[var]code" ✅ FIXED

Solution #3: Integrity Validation

Function: validate_translation_integrity()

Checks for missing placeholders after translation:

missing = validate_translation_integrity(text, placeholders)
if missing:
    print(f"⚠️ WARNING: Missing placeholders: {missing}")
    # Log, retry translation, or warn the user.

---

⚙️ Configuration

Default Behavior (v2.6.7)

In src/utils/config.py:

use_html_protection: bool = True  # Global default (Safe for Cloud APIs)

Recommended Config for Web Users:

{
    "use_html_protection": false,
    "translation_engine": "google",
    "verify_placeholders": true,
    "retry_failed_strings": true
}

---

🎯 Best Practices

✅ DOs for Web Google Translate

Action	Success Rate	Reason
Use Token Mode	✅ 100%	Foundation for web-based safety
Enable Integrity Validation	✅ 100%	Catches corruption early
Use Retry Logic	✅ 95%	Recovers from network glitches
Test Small Batches First	✅ 90%	Quick validation before full game

❌ DON'Ts for Web Google Translate

Action	Success Rate	Reason
Enable HTML Mode	❌ 0%	Web version doesn't support tags
Large Batch Sizes (>100)	❌ 15%	High risk of 429 Rate Limit
Long sessions without Proxy	❌ 5%	IP will eventually be flagged

---

🔧 Troubleshooting

Issue: PLACEHOLDER_CORRUPTED Errors

Symptoms:

[player_name] → [player OUTSIDE_NAME]
{color=#fff} → {color = #fff}

Solution:

1. Ensure Spaced Token Fix is enabled (v2.6.7+ performs this automatically).
2. If it persists, check the token regex in syntax_guard.py.

Issue: Rate Limiting (Error 429)

Symptoms:

• The process stops or skips many lines.
• "Too Many Requests" appears in logs.

Solutions:

1. Rotate Proxies: Use Settings > Proxy Manager.
2. Reduce Threads: Set Concurrent Threads to 1 or 2.
3. Increase Delay: Add 1-2 seconds between requests.

---

🚀 Improvement Roadmap (v2.6.8+)

Phase 1: Token Robustness

• A) Pattern Expansion: Add broader support for various spacing styles (e.g., "V AR 0").
• B) Fuzzy Matching: Use Jaro-Winkler or Levenshtein ratios to match corrupted tokens.
• C) Context-Aware Recovery: If "VAR" is expected but missing, scan nearby text for digits.

Phase 2: User Experience

• A) Auto-Detection: Automatically disable HTML protection if web-based endpoints are detected.
• B) Interactive Fix UI: Allow users to manually drag-and-drop missing placeholders into the translated text.

---

📊 Performance Comparison

Translation Speed (sec / 1000 lines)

Engine	Speed	Reliability	Cost
Web Google (Token)	120s	85% (Fixed)	Free
Paid Google Cloud API	15s	98% (HTML)	Paid
OpenAI (GPT-4o)	8s	99%	Paid
Local LLM (Llama 3)	60s+	95%	Free (needs GPU)

---

✨ Summary

If you are using Web Google Translate:

1. ✅ Set use_html_protection = false in Settings.
2. ✅ Update to v2.6.7+ for the automatic spaced token fix.
3. ✅ Keep Batch Size moderate to avoid rate limits.

Result: With the new recovery logic, RenLocalizer achieves a 95%+ success rate even on the free web-based Google Translate service.

---

Last updated: 8 Feb 2026 | RenLocalizer Documentation v2.6.7