CLI Dashboard¶
Overview¶
The llm-toll CLI provides terminal-based access to cost and usage statistics. It is installed automatically as a console script.
Commands¶
--stats -- Show Usage Statistics¶
# All projects summary
llm-toll --stats
# Filter by project
llm-toll --stats --project my_scraper
# Filter by model
llm-toll --stats --model gpt-4o
Output is a formatted table with color-coded costs:
- Green: < $0.01
- Yellow: $0.01 - $0.10
- Red: > $0.10
Colors are suppressed when the NO_COLOR environment variable is set.
--reset -- Reset Budget¶
Resets the accumulated cost for a project to zero. Requires --project.
--export csv -- Export Usage Logs¶
# Export all logs to stdout
llm-toll --export csv
# Export with filters
llm-toll --export csv --project my_scraper
llm-toll --export csv --model gpt-4o
# Export to a file
llm-toll --export csv --output report.csv
llm-toll --export csv --project my_scraper --output report.csv
CSV columns: project, model, input_tokens, output_tokens, cost, created_at.
--update-pricing -- Fetch Latest Pricing¶
Fetches the latest model pricing from the remote source and caches it locally. See Pricing Updates.
--dashboard -- Launch Web Dashboard¶
Starts a browser-based dashboard at http://127.0.0.1:8050 (or the specified port). See Web Dashboard.
Global Options¶
| Option | Description |
|---|---|
--version |
Show version and exit |
--db PATH |
Path to the SQLite database (default: ~/.llm_toll.db) |
--store-url URL |
Store URL for PostgreSQL (e.g., postgresql://user:pass@host/db) |
--project NAME |
Filter by project name |
--model NAME |
Filter by model name |
--output FILE |
Output file for --export (default: stdout) |
--port PORT |
Port for --dashboard (default: 8050) |
Examples¶
# View all projects
llm-toll --stats
# View per-model breakdown for a project
llm-toll --stats --project my_app
# View per-project breakdown for a model
llm-toll --stats --model gpt-4o
# Reset a project's budget counter
llm-toll --reset --project my_app
# Export to CSV file
llm-toll --export csv --output monthly_report.csv
# Use PostgreSQL backend
llm-toll --stats --store-url postgresql://user:pass@host/llm_costs
# Use a custom SQLite database
llm-toll --stats --db /path/to/my.db