Budget Mode
Transparent cost control for AI operations. Know exactly what you're spending before you spend it.
Word-Based Pricing
How It Works
Instead of abstract credits or tokens, Lexic shows costs in words processed. This gives you an intuitive sense of what operations cost relative to the content you're working with.
See Cost First
Preview the word cost before running any operation
Real-Time Balance
Your word balance updates instantly after each operation
No Surprises
Operations are blocked if they would exceed your balance
Operation Costs
Each AI operation has a fixed cost in words. Here's what to expect:
| Operation | Words | Description |
|---|---|---|
| Entity Extraction | ~1,000 | Extract people, organizations, and concepts from your notes |
| Embeddings | ~100 | Generate semantic vectors for search and connections |
| Summarization | ~300 | AI-generated summaries of your content |
| Deep Analysis | ~1,000 | Comprehensive analysis of content and connections |
| Chat | ~200 | AI conversation about your knowledge |
* Word counts are approximate and may vary based on input size.
Budget Controls
Balance Tracking
Each lexicon has its own word balance. You can see your current balance at any time and track how it changes with each operation.
Word Grants
Words are added to your balance through grants, which can be:
- •Full allocation (e.g., monthly allowance)
- •Partial top-up (add specific amount)
Word Holds
Temporarily freeze words to prevent usage. Useful for:
- •Reserving budget for important projects
- •Pausing team usage during audits
Word Refunds
If an operation fails or is cancelled, words are automatically refunded. All refunds are tracked in the audit log for transparency.
FIFO Consumption
Words are consumed in the order they were granted (oldest first). This ensures granted words are used before they expire.
30-Day Expiration: Granted words expire after 30 days if unused. This encourages regular usage and helps us maintain predictable infrastructure costs. You'll be notified before words expire.
Team Budget Management
Teams OnlyOrganization Word Pool
Teams share a central word pool across all team lexicons. This makes it easy to manage costs at the organization level rather than tracking individual allowances.
Shared Pool
All team lexicons draw from the same word balance
Tier-Based Limits
Support word limits based on your subscription tier
Bulk Operations
Grant, hold, or refund words across multiple lexicons at once
Personal vs Team Words
Team members' personal lexicons use their own individual words, completely separate from the organization pool. This ensures personal work never affects team budgets (and vice versa).
Backlog System
Safety NetYour work is never lost. Even if you run out of words, your notes are saved and queued for processing.
How It Works
When words run out, AI operations (entity extraction, embeddings, etc.) are backlogged instead of failing. Your content is saved immediately—only the AI processing is deferred.
While Backlogged
- ✓Notes are saved and accessible
- ✓You can view and edit content
- ✓Status shows "Pending Processing"
- ○AI features unavailable until processed
When Words Are Added
- ✓Backlog is automatically processed
- ✓Oldest items are processed first (FIFO)
- ✓No manual intervention required
- ✓Processing happens in the background
Backlog Status Tracking
You can always see the state of your backlog:
Cost-Saving Tips
Do
- ✓Write comprehensive notes—longer notes extract more value per operation
- ✓Batch related content into single notes when appropriate
- ✓Monitor your balance regularly to avoid surprises
Avoid
- ✗Creating many tiny notes for content that belongs together
- ✗Running deep analysis on content that only needs basic processing
- ✗Letting words expire—use them or they're gone