Reading the usage page
Understand the charts, totals, and bot table on the Usage page.
The Usage page shows how many requests and credits your workspace consumed over a selected date range.
Pick a date range
Use the date range controls or quick presets:
- 1D — today.
- 7D — the last 7 days.
- 30D — the last 30 days.
Click Apply after changing dates.
Top metrics
- Total Requests — number of requests in the selected range.
- Total Credits — credits consumed in the selected range.
- Credits by type — visible when Source is All, splitting Chat and Training credits.
Usage over time
The chart plots daily request/credit activity. Spikes usually mean a popular widget page, a large training run, or API automation.
By Chatbot table
Each row shows:
- Bot name and id.
- Relation to the current workspace.
- Total requests.
- Total credits.
If a bot name is missing, Pegasus shows "Unknown Bot" while still preserving the usage record.
Source filter
The Source filter narrows usage to where requests came from.
Source values
| Source | Includes |
|---|---|
| Direct | Logged-in workspace chat. |
| Widget | Visitor conversations from the embed widget. |
| API | Personal Access Token/API usage. |
| All | Every source, plus training credits where applicable. |
Training credits and filtering
When you filter by a specific source, Pegasus does not include training credits. Switch to All to see every source and the chat/training breakdown.
When to use each filter
- Use Direct when investigating your own workspace chat.
- Use Widget when a public website has sudden traffic.
- Use API when checking automation or external clients.
- Use All for the monthly total picture.
Credits breakdown
The Credits card can show a breakdown by type when the Source filter is All.
Chat credits
Chat credits include:
- Workspace conversations.
- Widget visitor messages.
- API/PAT messages.
- Extra context and response tokens.
Training credits
Training credits include document processing work such as parsing and embedding. They are linked to bot knowledge updates, not to a visitor or chat session.
Why numbers differ from request count
One request can consume multiple credits. A long response, a large context window, or Max Mode can increase token usage.
Use the breakdown for decisions
- If Chat dominates, inspect popular bots and source filters.
- If Training dominates, review document size and retraining frequency.
- If Widget dominates, check allowed origins and public-site traffic.
Relation labels
Relation labels explain why a usage row appears in the current workspace.
| Relation | Meaning |
|---|---|
| Personal owner | You own the bot in Personal workspace. |
| Team private owner | You own a private bot inside the team. |
| Team shared owner | You own a team bot shared with members. |
| Shared viewer | You used a bot shared by another team member. |
| Owner managed | You are viewing a team bot with owner-level management context. |
| Widget owner | Usage came from your bot's embed widget. |
| Unknown | Pegasus could not classify the relation for this row. |
Relations keep team usage fair. A shared bot can be owned by one person, used by another, and still belong to the team workspace context.
If a relation looks unexpected:
- Switch to the workspace where the bot lives.
- Check whether the bot is shared with team.
- Check whether usage came from Widget or API source.
- Ask the team owner if an owner-managed bot appears unexpectedly.
Team usage views
Team owners see additional usage tabs when they are in a team workspace.
My Usage
Shows usage attributed to you in the current workspace context.
Team Members
Shows usage grouped by member. Use this to understand which users are consuming team quota.
Team Bots
Shows usage grouped by bot. Use this when a shared or owner-managed bot is consuming unexpectedly high quota.
Who can see team views
Team owner role is required for team-wide member/bot usage views. Regular members see their own usage context.
The selected date range and Source filter apply to all tabs.