Data Retention
Understanding how TeaL;DR stores and manages your server's data.
Overview
TeaL;DR automatically manages data retention to balance searchability with privacy and storage efficiency. All retention settings are configurable by server administrators.
Default Retention Policy
- •Messages: Automatically deleted after 30 days
- •Embeddings: Deleted with associated messages
- •Analytics: Aggregated data retained for 90 days
- •Configuration: Retained until bot removal
Configuring Retention Period
Set Custom Retention Period
Administrators can adjust the message retention period between 7 and 90 days:
/settings action: Set Retention Period days: 60Minimum: 7 days
Maximum: 90 days
Default: 30 days
View Current Retention
Check your current retention period setting:
/settings action: View Retention Period7-14 Days
Short-term retention
- • Recent discussions only
- • Minimal storage
- • High privacy
30 Days
Recommended (Default)
- • Balanced approach
- • Good searchability
- • Reasonable storage
60-90 Days
Long-term archival
- • Extended history
- • More storage needed
- • Comprehensive search
What Gets Stored
Message Data
For each indexed message, we store:
- Content: Full message text
- Metadata: Author ID, channel ID, timestamp
- Embeddings: AI-generated semantic vectors (768 dimensions)
- Context: Thread information, reply references
Storage per message: ~2-5 KB (text + embeddings)
Analytics Data
Aggregated statistics stored separately:
- Message counts per user/channel
- Activity patterns and trends
- Topic distributions
- Engagement metrics
Note: Analytics are anonymized and aggregated
Configuration Data
Server settings and preferences:
- Excluded channels list
- Retention period setting
- Bot personality preference
- Permission configurations
Retention: Kept until bot is removed from server
Automatic Deletion
How It Works
- Daily Cleanup: Bot runs automated cleanup every 24 hours
- Age Check: Messages older than retention period are identified
- Deletion: Messages and embeddings are permanently deleted
- Analytics Update: Aggregated stats are updated
⏰ Timing
Automatic deletion runs at 00:00 UTC daily. Messages are deleted based on their creation timestamp, not when they were indexed.
Manual Data Management
Clear Channel Data
Delete all indexed data from a specific channel:
/settings action: Clear Channel Data channel: #old-channelClear All Server Data
Delete all indexed data for your entire server:
/settings action: Clear All Data confirm: yesBot Removal
When you remove the bot from your server:
- New message indexing stops immediately
- Existing data is retained for 30 days
- After 30 days, all data is permanently deleted
- Re-adding the bot within 30 days restores configuration
Storage Estimates
Typical Server Storage
| Server Size | Messages/Day | 30-Day Storage | 90-Day Storage |
|---|---|---|---|
| Small | 100-500 | ~50-250 MB | ~150-750 MB |
| Medium | 500-2000 | ~250 MB-1 GB | ~750 MB-3 GB |
| Large | 2000-10000 | ~1-5 GB | ~3-15 GB |
| Very Large | 10000+ | ~5+ GB | ~15+ GB |
Note: Estimates include message content, embeddings, and metadata. Actual storage may vary based on message length and attachment frequency.
Retention Best Practices
Choosing the Right Period
- •Active Servers: 30 days provides good balance for most use cases
- •Archival Needs: 60-90 days for servers requiring longer history
- •Privacy Focus: 7-14 days for maximum privacy and minimal storage
- •High Traffic: Shorter periods help manage storage on very active servers
Regular Maintenance
- Review retention settings quarterly
- Monitor storage usage with
/stats - Adjust based on server activity patterns
- Clear old channel data when channels are archived
- Document retention policy in server rules
Compliance Considerations
- Ensure retention period complies with your privacy policy
- Consider GDPR/data protection requirements
- Document data retention in Terms of Service
- Provide clear deletion procedures for users
🔒 Privacy Guarantee
All data is isolated per server. Your server's data is never shared with other servers or used for training AI models. Data deletion is permanent and cannot be recovered.