# Reporting: Words, Images, Tokens & Cost

> **What it does:** A dedicated report of everything the AI produced — words generated, images created, tokens used, and an **estimated cost per post** — so you always know what your content actually costs.
>
> *Applies to Antradus AI Pro v2.0.0 · Last updated 2026-07-24*

## Where it lives

**Antradus AI → Reporting.** All data stays in **your own database** — nothing is sent anywhere.

<!-- screenshot: reporting-1 — the reporting table with per-post rows -->

## What you see

| Column | Meaning |
|---|---|
| **Post** | Which article the usage belongs to (bulk posts included) |
| **Words** | Words generated (counts correctly for Chinese/Japanese/Korean too) |
| **Images** | Images generated for that post |
| **Tokens** | Exact token counts reported by your provider |
| **Est. cost** | Tokens × your model prices + images × your image price |

## Making the cost estimate accurate

Prices are **pre-filled and fully editable**:

1. On the Reporting page, find the model-prices section.
2. Enter your provider's current per-token (per-million) rates for the models you use, and your per-image price.
3. Costs recalculate from your numbers.

This matters for agencies: set your *real* rates once and the report becomes a true cost sheet you can put next to client invoices.

## Agency use

- Combine with the **Bulk Publishing → 📊 Reports** CSV export: batch report for the client, cost report for you — margin visible per project.
- Free models (OpenRouter ★ FREE) show as zero cost, correctly.

## Tips & gotchas

- Usage is attributed to a post when there is one; utility calls (e.g. topic suggestions) are tracked too.
- Estimates are estimates — providers bill in their own dashboards; treat this as your management view, their invoice as the truth.
- Old rows can be purged from the page if you want a clean slate.