Token Report Plugin
Adds a Token Usage section to the end-of-session report: model, estimated cost, input/output split, per-query averages, duration, and a per-query bar chart.
Enabled in the shipped default config. Printed by Console Report.
Configuration
plugins:
- module: codumentor.plugins.token_report
class: TokenReportPlugin
No plugin args. Requires session_report: true and Console Report.
Usage
No web or chat UI. After a print-mode (-p) run the session report includes:
Token Usage
Model: gpt-4o Cost: $0.0124
Total Tokens: 15.2K Queries: 8
Input: 12.4K ( 81.6%) ████████████████░░░░
Output: 2.8K ( 18.4%) ████░░░░░░░░░░░░░░░░
Per Query Avg: Total: 1900.0 Input: 1550.0 Output: 350.0
Duration: 47.3s
Per-query token usage (bars show relative token counts)
██ ████ ██ ██ ████ ██ ░░ ██
Cost is N/A when the model does not report an estimate. The per-query chart appears only when there were two or more queries.
Notes
- Print-mode will not emit any session report unless this plugin and Console Report are both loaded.
- There is no in-chat token display.