# Check AI prompt logs and fix errors

> See every run of an AI prompt, how many tokens it used and why a run failed, and track your daily AI token usage.

Source: https://tadabase.com/docs/ai-features/check-ai-prompt-logs  
Status: Verified Nov 17, 2026. Checked against the current product

Every run of an AI prompt is recorded on that prompt's **Logs** tab, and the AI Prompts screen shows your token use per day. Use them to confirm a prompt ran, see how many tokens it used, and find out why a run failed.

## Open a prompt's logs

1. **Open the prompt** Go to **Automation** > **AI Prompts** and click the prompt.
2. **Open Logs** Click the **Logs** tab. It appears once the prompt has been saved.
3. **Load new runs** Runs are written in the background, so a run that just finished can take a moment to appear. Click **Refresh** to load it.

Runs are listed newest first, a page at a time; use the page controls next to **Refresh** to go back further. The list has no search or filters.

| Column | What it shows |
| --- | --- |
| **Date** | When the prompt ran. |
| **User Prompt** | The user prompt that was sent. |
| **Total Tokens** | Tokens used by the run. A failed run shows 0. |
| **Status** | `success` or `error`. For an error, hover over the icon next to the status to read the error message. |

Runs from forms, components and workflows are logged, and so are **Test Playground** runs of a saved prompt. If the list is empty you see No AI prompt logs were found.

![The Logs tab of an AI prompt, listing runs with their date, user prompt, total tokens and status](https://tadabase.com/docs-shots/ai-features-prompt-logs.webp)

## See your daily token usage

On the **AI Prompts** screen, click **View Daily Usages**. The **AI Usage Logs** window lists the **Date** and **Total Tokens** for up to the last 50 days, with the overall total at the top.

- The numbers are for the whole account, not only this app: every app on the account shares the daily limit.
- The button shows once the app has at least one AI prompt.
- You need to be the account owner or have access to the app's settings. Otherwise the window shows You do not have permission to access this data.

## Common errors

| Message | Where you see it | What it means and what to do |
| --- | --- | --- |
| `Daily AI Prompt Tokens Limit Reached` | **Logs** tab (error tooltip) | The account used its daily AI token allowance. Runs work again the next day, or after you move to a plan with a higher limit. Check use with **View Daily Usages**. |
| You have reached your daily AI usage limit. Please try again tomorrow. | **Test Playground** | Same as above, for a test run. |
| A plan message instead of the AI Prompts list, or when you save or test | **AI Prompts** screen, **Save**, **Test Playground** | Your plan does not include AI prompts. Use the upgrade button on the screen or change plans. |
| No **Allow Attachment** switch on the prompt form | **General** tab | Your plan does not include file attachments in AI prompts. |
| We couldn't read the attached file. Check the link is a PDF or image under 50 MB. | **Test Playground** | The **Attachment URL** could not be read. Use a link to a PDF or an image under 50 MB. |
| The Response schema for this prompt is invalid. Please review the Response tab. | **Test Playground** | A response on the **Response** tab is set up in a way the AI service rejects. Check each response's **Name**, **Type** and **Allowed Responses**. |
| The AI service blocked this response for safety reasons. Try adjusting your prompt. | **Test Playground** | The AI service refused to answer. Reword the prompt or the test values. |
| The AI service is temporarily rate-limited. Please try again in a moment. | **Test Playground** | Too many requests reached the AI service at once. Wait and try again. |
| The AI service took too long to respond. Please try again. | **Test Playground** | The run timed out. Try again. |
| User prompt and system prompt are required to run a test | **Test Playground** | Fill in both **User Prompt** and **System Prompt** on the **General** tab. |

> **Note:** On the **Logs** tab, a failed run's tooltip shows the error as the AI service returned it, which can be more technical than the **Test Playground** messages above. To see a friendly message, run the same values in the **Test Playground**.
