Babel v2.5.7-beta
Updating Babel...
Installing

Babel will restart briefly. The page will reconnect automatically.

All services and settings preserved.
Update failed

Babel is still running the previous version ().

Latest
What's new
Performance improvements and bug fixes.
Total Jobs
Translated
Existing / Skipped
Avg Translation Time

Activity Feed

Live Sync
Title / Video
Source
Status
Lines
Duration
Time

AI Usage & Queue

Monitor daily AI request limits, token accounting, estimated costs, and deferred translation queues.

Requests Today
UTC Day Window (00:00 reset)
Daily Request Limit
Active:
Remaining
Estimated Cost Today
Estimated Token Cost
Provider Status

Daily Request Limit

Optional safety limit for AI requests per day. Set to 0 for unlimited.

Unlimited
Set 0 for unlimited. Babel gracefully defers jobs when limit is reached.

Today's Usage by Provider

Capacity Estimate

Based on historical averages
Avg AI calls
Avg cost
Estimated episodes remaining today
Based on remaining request capacity
Unlimited
Sample: translated jobs ( total calls)

Deferred Jobs Queue

Processed in FIFO order as capacity becomes available

AI Translation Engine

Choose your AI translation provider and configure parameters.

DeepL provides dedicated neural translation without general LLM prompt parameters.

Contextual Escalation

Automatically retry stubborn lines with a stronger model and previous/next dialogue context.

Same as Primary ()

Speed & Batch Parameters

Subtitle lines sent per AI request. Larger batches use fewer requests and more context.
AI requests processed in parallel for each video.
Videos that can translate at the same time.
0 keeps provider defaults (60 seconds for Ollama). Set up to 3600 seconds for slower inference. Applies to translation, QA, and recovery requests.
Additional preferences added to Babel's built-in translation instructions. Core formatting and subtitle integrity rules remain enforced.
Optional. One rule per line: English term = Target translation. Use this to force the AI to keep specific names or places intact.
Rate Limit Guidelines (Google AI Studio Official):
• Free Tier (15 RPM limit): Set Simultaneous Videos to 1 to stay safely within Google's 15 requests/minute free threshold.
• Pay-As-You-Go / Tier 1 (300+ RPM limit): Set Simultaneous Videos to 3 (or higher) to translate entire seasons in parallel at ultra-speed.
Tip: Flash-Lite and Flash start at 300 RPM on Tier 1 (scaling up to 1,000+ RPM as usage grows), while Pro models have stricter limits (150 RPM).
OpenAI Rate Limit Guidelines:
• Tier 1+ Account: Set Simultaneous Videos to 2. GPT-4o Mini is fast and cost-effective for subtitle translations.

Features & Modules

Enable or disable subtitle extraction and processing modules.

SDH & Noise Cleaner

Automatically cleans SDH sound descriptions, background noise, and musical notes while maintaining flawless sync.

Extract Embedded Target Subtitles

If target language already exists inside the video, extracts it directly as an external .[lang].srt file.

Extract Embedded Source Subtitles

Extracts embedded source subtitles from MKV/MP4 containers (Priority 1 — highest sync confidence). The best available track is selected automatically.

Auto-Repair Corrupted Subtitles

Inspects existing subtitles and re-translates if severe sync or line corruption is detected.

Strict Source Sync Lock

Locks start and end timestamps directly to the source subtitle timing.

Target Languages & Priority

Select which languages to translate to and set their priority order.

Media Root Folders

Specify the paths inside the Babel container where your TV shows and Movies reside.

Default: /tv (or /media/tv)
Default: /movies (or /media/movies)

Docker Path Mapping (Sonarr/Radarr Webhooks)

If Sonarr and Babel use different volume mounts, map the remote path to the local path.

Integrations & Pipeline Mode

Choose how Babel handles subtitles: Hybrid Mode with Bazarr, or Pure AI Direct.

Hybrid Mode

Parallel Prep + Bazarr Coordination

Babel checks for human subtitles through Bazarr while preparing the AI fallback. If no target subtitle is available when preparation completes, translation starts immediately.

Pure AI Mode

AI First

Babel skips Bazarr searches and immediately uses the best available source subtitle for AI translation.

Bazarr Integration Parameters
Hybrid mode: With Grace Period set to 0, Bazarr searches concurrently while Babel prepares the AI fallback. A configured Grace Period delays automatic AI translation.
Grace Period before AI fallback

New Sonarr and Radarr items wait for human subtitles before AI translation. 0 days and 0 hours keeps the current behavior. Manual translation bypasses the wait. Changes apply from each waiting item's original start.

Jellyfin Media Server

Instantly notify Jellyfin when a new subtitle is ready

Plex Media Server

Scan the matching Plex library when a new subtitle is ready

Only needed when Plex sees your media under a different path than Babel. Leave blank if both use the same paths.

Webhooks & Arr Automation

Sniper-Webhook URLs to trigger Babel automatically when Sonarr or Radarr finishes a download.

Sonarr Webhook Endpoint POST
Use an address that Sonarr can reach.
Same Docker network only:
Radarr Webhook Endpoint POST
Use an address that Radarr can reach.
Same Docker network only:
How to set up in Sonarr / Radarr:
1. Go to Settings → Connect → click + → select Webhook.
2. Name: Babel.
3. Enable On File Import and On File Upgrade.
4. Paste the webhook URL shown above.
5. If BABEL_WEBHOOK_SECRET is configured, add custom header X-Webhook-Secret: YOUR_SECRET (preferred) or append ?secret=YOUR_SECRET.
6. Set Method to POST.
7. Click Test (will show green checkmark) → Save.

System & Updates

Manage Babel system version and in-app updates.

Current Version
Checking for updates...
Updating...
Updated to
Update failed
Unable to check for updates
Update available ·
Up to date
Updating...
Please wait, Babel is restarting. ()
Update failed

Babel was safely rolled back or remained on version . All settings and database state are preserved.

Status
Duration
Lines Processed
Sync Drift & Dropped

Step-by-Step Execution Log

AI Usage

What's New in

Release details & changelog from GitHub

View full release on GitHub