> For the complete documentation index, see [llms.txt](https://docs.qolaba.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.qolaba.ai/common-issues.md).

# Common Issues

<details open>

<summary><mark style="color:orange;"><strong>Generation failed — what should I do?</strong></mark></summary>

* Check your remaining credit balance in the Overview tab. Insufficient credits will cause generation failures.
* Verify your internet connection is stable.
* If the model is temporarily unavailable, try switching to an alternative model with similar capabilities.
* Shorten or simplify your prompt if the request may have triggered a content filter.
* Refresh the page and retry. Temporary server-side issues often resolve within minutes.

</details>

<details open>

<summary><mark style="color:orange;"><strong>My credits don't match what I expected</strong></mark></summary>

* Credit consumption is based on input + output token length for LLMs. Longer prompts and longer responses cost more.
* For images, total cost = credits per image × number of generations.
* Knowledge Base uploads consume credits based on file size.
* Check the Credit Utilization section in the Overview tab for a full breakdown by tool and date.
* If you believe there is a genuine credit discrepancy, contact <support@qolaba.ai> with the date, tool used, and expected vs actual credit deduction.

</details>

<details open>

<summary><mark style="color:orange;"><strong>My credits don't match what I expected</strong></mark></summary>

* Credit consumption is based on input + output token length for LLMs. Longer prompts and longer responses cost more.
* For images, total cost = credits per image × number of generations.
* Knowledge Base uploads consume credits based on file size.
* Check the Credit Utilization section in the Overview tab for a full breakdown by tool and date.
* If you believe there is a genuine credit discrepancy, contact <support@qolaba.ai> with the date, tool used, and expected vs actual credit deduction.

</details>

<details open>

<summary><mark style="color:orange;"><strong>Invited member cannot access the workspace</strong></mark></summary>

* Confirm the member signed in using the exact email address they were invited with.
* Check the member's status in the Members tab — ensure it shows 'Active', not 'Revoked'.
* Verify the correct workspace is assigned to the member.
* Ask the member to log out and log back in to refresh their session permissions.

</details>

<details open>

<summary><mark style="color:orange;"><strong>Image or video generation is slow</strong></mark></summary>

* Premium models (Gemini Nano Banana Pro, Runway Gen4 Turbo) have higher processing times than lighter models.
* Generating multiple outputs simultaneously (e.g., 8 images) takes longer than single generations.
* High-quality and detail settings increase processing time. Reduce quality settings for faster results during drafting.
* Platform-wide load spikes can temporarily increase queue times. Try again after a few minutes.

</details>

<details open>

<summary><mark style="color:orange;"><strong>Knowledge Base responses are inaccurate or missing context</strong></mark></summary>

* Ensure the Knowledge Base is actively attached to the current chat (a file icon should appear in the chat interface).
* Verify the relevant documents are included in the Knowledge Base and have been successfully processed.
* Large files may index partially. Try splitting very large documents into smaller chunks.
* Remove outdated or conflicting documents that may be overriding current information.

</details>

<details open>

<summary><mark style="color:orange;"><strong>Audio generation sounds unnatural</strong></mark></summary>

* Use Pro TTS model for final output — Flash TTS is optimized for speed, not naturalness.
* Add style instructions to guide delivery: 'Speak slowly and warmly' or 'Professional and authoritative'.
* Use proper punctuation in your script — commas and full stops control pacing significantly.
* Break very long paragraphs into shorter sentences for more natural pacing.

</details>