Skip to main content

​
Rate Limits & Quotas

X21 has several limits to ensure optimal performance and manage costs. Understanding these helps you work within constraints.

​
Token Limits

​
Conversation Limit

200,000 tokens per conversation Includes:
  • All your messages
  • All AI responses
  • Tool definitions (background)
  • System messages
  • Attached files (~100KB per PDF page estimate)

​
Per-Response Limits

  • Output: 32,000 tokens reserved
  • Thinking: 1,600 tokens reserved
  • Combined: Up to 33,600 tokens per AI response

​
What Are Tokens?

Tokens are units of text:
  • ~4 characters = 1 token
  • “Hello” = 1 token
  • “spreadsheet” = 2 tokens
  • Full conversation history counts

​
Token Counter

Status bar shows real-time usage: 
📊 Tokens: 45,230 / 200,000

​
When Limits Are Reached

At 200,000 tokens, X21 automatically:
  1. Compacts conversation (summarizes old messages)
  2. Preserves recent context
  3. Continues without interruption
Best practice: Start new conversations for unrelated tasks.

​
File Limits

​
PDF Attachments

100 pages total per request Examples:
  • 1 file × 100 pages = OK
  • 2 files × 50 pages each = OK
  • 5 files × 30 pages each = Exceeds limit

​
Image Files

No explicit limit, but:
  • Large images increase processing time
  • Multiple images count toward token usage
  • Recommended: Compress large images

​
File Types

Supported:
  • PDFs (up to 100 pages total)
  • PNG, JPG, JPEG, GIF, WEBM
Not supported:
  • Excel files (use file operations instead)
  • Word documents
  • Other formats

​
Query Limits

​
Recent Chats

Max 50 conversations per history query Configurable range: 1-50

​
Search Results

Max 100 results per search query Configurable range: 1-100

​
Messages Per Conversation

No hard limit, but:
  • Very long conversations may compact
  • Performance may degrade beyond 200k tokens

​
Rate Limiting

​
Anthropic API Limits

X21 uses Claude API, which has rate limits: If rate limited:
  • Error message appears
  • Wait time indicated (typically 30-60 seconds)
  • Retry automatically or manually
Common causes:
  • Multiple rapid requests
  • Large file processing
  • Peak usage times

​
Recovery

⚠️ Rate limit reached. Please wait 30 seconds.
Actions:
  1. Wait the indicated time
  2. Retry your request
  3. Contact support if persistent

​
Best Practices

​
Managing Tokens

Start fresh:
  • New conversation for new tasks
  • Don’t mix unrelated work
  • Clear separation reduces token usage
Monitor usage:
  • Watch the token counter
  • Compact before hitting 200k
  • Use concise prompts when possible

​
File Attachments

Optimize PDFs:
  • Extract relevant pages only
  • Compress before attaching
  • Split large documents
Image efficiency:
  • Resize to reasonable dimensions
  • Compress without losing quality
  • Screenshot only necessary portions

​
Avoiding Rate Limits

Pace requests:
  • Don’t rapid-fire multiple requests
  • Let responses complete
  • Batch operations when possible
Off-peak usage:
  • Fewer rate limits during off-peak hours
  • Plan large operations accordingly

​
Error Messages

​
Token Limit Errors

⚠️ Conversation approaching token limit
Action: Start new conversation soon 
⚠️ Request too large. Try removing attachments or starting a new conversation.
Action: Reduce file attachments or start fresh

​
Rate Limit Errors

⚠️ Rate limit reached. Please wait 30 seconds.
Action: Wait and retry 
⚠️ Service overloaded. Please try again in a moment.
Action: Brief wait, then retry

​
File Limit Errors

⚠️ PDF attachments exceed 100 pages total
Action: Remove pages or split documents

​
Quotas by User Type

All X21 users have the same limits:
  • 200,000 tokens per conversation
  • 100 PDF pages per request
  • Shared Anthropic API rate limits
Enterprise deployments may have custom limits - contact your administrator.