Monthly Summary
The monthly summary card at the top of the Usage page shows your current billing cycle's key metrics at a glance.
| Metric | Description |
|---|---|
| Total credits used | Credits consumed across all services this billing cycle. |
| Credits remaining | Current credit balance (subscription + purchased credits). |
| Gateway requests | Number of inference requests through the Creor Gateway. |
| Cloud agent runs | Number of cloud agent invocations this cycle. |
| Total tokens | Sum of input and output tokens across all models. |
| Billing period | Start and end dates of the current cycle. |
The summary updates in near real-time. A small delay (up to 5 minutes) may occur between a request and its appearance in the summary.
Cost by Model
The cost breakdown shows how your credits are distributed across different LLM models. This helps you identify which models consume the most resources and whether switching to a different model could reduce costs.
| Model | Requests | Input Tokens | Output Tokens | Credits |
|---|---|---|---|---|
| Claude Sonnet 4 | 1,234 | 4.2M | 380K | 142.50 |
| Claude Haiku 3.5 | 856 | 2.1M | 190K | 18.20 |
| GPT-4o | 423 | 1.8M | 210K | 98.30 |
| GPT-4o mini | 2,100 | 3.5M | 420K | 12.80 |
| Claude Opus 4 | 45 | 890K | 120K | 67.40 |
The table is sortable by any column. Click on a model name to drill down into daily usage for that specific model.
Tip
Token Usage
The token breakdown shows input vs output token consumption. This matters because most providers charge differently for input and output tokens.
Input Tokens
Input tokens include your message, system prompt, tool definitions, project context, and any code the agent reads. Input tokens are typically 3-10x more than output tokens because the agent reads a lot of code before producing a response.
Output Tokens
Output tokens include the agent's response text, tool call arguments (file edits, shell commands), and thinking tokens (for models that support extended thinking).
Reducing Token Usage
- Use project instructions (CREOR.md) to give the agent context upfront, reducing the need for exploratory tool calls.
- Scope your requests to specific files or directories when possible.
- Use the Plan agent for read-only analysis -- it generates plans without expensive edit/bash cycles.
- Enable session compaction to reduce token accumulation in long conversations.
- Close and start new sessions for unrelated tasks instead of continuing a long thread.
Daily Charts
Interactive charts show your usage patterns over time. Available views include:
- Credits per day: bar chart showing daily credit consumption.
- Requests per day: line chart showing daily request volume.
- Tokens per day: stacked area chart showing input vs output tokens.
- Cost by model per day: stacked bar chart breaking down daily cost by model.
Reading the Charts
Hover over any data point to see exact values. Click and drag to zoom into a date range. Use the time range selector (7 days, 30 days, 90 days, custom) to adjust the view.
Spikes in usage often correlate with specific events -- a large refactoring session, a batch of cloud agent runs, or a new team member onboarding. Use the charts to understand these patterns and set appropriate spending limits.
Usage History
The history table at the bottom of the page shows every billable event in reverse chronological order. Filter by date range, model, API key, or event type.
| Column | Description |
|---|---|
| Timestamp | When the request was made. |
| Type | Gateway inference, cloud agent compute, or cloud agent inference. |
| Model | Which LLM model was used. |
| API Key | Which key authenticated the request (name, not the key itself). |
| Input tokens | Number of input tokens consumed. |
| Output tokens | Number of output tokens consumed. |
| Credits | Credits charged for this event. |
Export the history table as a CSV file for external analysis or expense reporting. Click the "Export CSV" button in the top right of the table.
API Access
Access usage data programmatically for custom dashboards, alerts, or integration with your internal billing systems.