> For the complete documentation index, see [llms.txt](https://yvechat.gitbook.io/wp/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://yvechat.gitbook.io/wp/getting-started/troubleshooting.md).
# Troubleshooting
Each section describes a symptom, the usual cause, and the steps that resolve it.
Common states and the step that clears each
***
### YveChat Engine unreachable
**Symptom.** The workspace reports that it cannot reach YveChat Engine.
**Usual cause.** YveChat Engine has stopped, or was never started.
**Steps.**
1. Confirm YveChat Engine is running on your computer.
2. Open its loopback address directly in a browser. The local approval page confirms it is alive.
3. Restart YveChat Engine if that address fails to load.
4. Check whether a firewall or security tool is blocking local connections.
5. Confirm nothing else on your machine occupies the same port.
***
### Browser permission denied
**Symptom.** The workspace cannot connect, and no permission prompt appears.
**Usual cause.** The local-network permission was declined, and the browser remembers the decision.
**Steps.**
1. Open your browser's site settings for `app.yvechat.xyz`.
2. Find the local network permission and set it to allow.
3. Reload the workspace.
Chrome gates connections from public sites to loopback addresses behind this permission, and the restriction has tightened across releases.
Source: [Chrome Local Network Access](https://developer.chrome.com/blog/local-network-access)
On Safari, loopback behavior has varied across versions, and a WebKit bug covering it has been open since 2017.
Source: [WebKit 171934](https://bugs.webkit.org/show_bug.cgi?id=171934)
Using the local YveChat Engine address directly works when a browser blocks the hosted path.
***
### Ollama unavailable
**Symptom.** The workspace connects to YveChat Engine, and language models fail.
**Usual cause.** Ollama has stopped, or was never started.
**Steps.**
1. Open `http://127.0.0.1:11434` in a browser. A confirmation message means Ollama is running.
2. Start Ollama if that address fails.
3. Restart YveChat Engine so it rediscovers the runtime.
4. Reinstall Ollama if the service refuses to start. See [Install Ollama](/wp/getting-started/install-ollama.md).
***
### Empty model list
**Symptom.** YveChat Engine is connected, and the model selector shows nothing.
**Usual cause.** Ollama is running with no models installed, or a newly pulled model has yet to be discovered.
**Steps.**
1. Use the refresh control under the model selector. Discovery is dynamic, and a refresh picks up newly installed models.
2. Confirm at least one model is installed through Ollama.
3. Download one if the list is genuinely empty. See [Choose Your First Model](/wp/getting-started/choose-your-first-model.md).
4. Check any capability filter that is active, since a filter hides models lacking that capability.
***
### Expired pairing request
**Symptom.** The six-digit code is rejected, or the approval page reports the request has expired.
**Usual cause.** Pairing codes expire after a short period.
**Steps.**
1. Return to the hosted workspace.
2. Request a new pairing code.
3. Enter the new code promptly on the local approval page.
4. Verify the displayed origin again before approving. See [Pair Your Browser](/wp/getting-started/pair-your-browser.md).
***
### Insufficient memory
**Symptom.** A model fails to load, generation stops partway, or the machine becomes unresponsive.
**Usual cause.** The model needs more memory than the machine has available.
**Steps.**
1. Close other demanding applications.
2. Choose a smaller model, or a more heavily quantized version of the same one.
3. Reduce the configured context length. Memory scales with parallel requests multiplied by context length.
4. Avoid running image generation and language generation at the same time.
Source: [Ollama FAQ](https://docs.ollama.com/faq)
***
### Slow first response
**Symptom.** The first reply from a model takes far longer than later replies.
**Usual cause.** The model is loading into memory. This is expected behavior.
**Steps.**
1. Wait for the first response to complete. Later replies in the same session start faster.
2. Expect the first reply after switching models to take longer when Ollama needs to load the selected model. Whether the previous model remains loaded depends on Ollama's configuration and available memory.
3. Confirm which hardware is handling generation if every response stays slow. YveChat Engine reports the detected acceleration, and processor-only generation is considerably slower.
4. Consider a smaller model on machines without a supported GPU.
***
### ComfyUI or image model unavailable
**Symptom.** Image Studio reports no available models, or image generation fails.
**Usual cause.** ComfyUI is absent, stopped, or has no image models installed.
**Steps.**
1. Confirm ComfyUI is installed and running on your machine.
2. Restart YveChat Engine so it rediscovers the image registry.
3. Confirm at least one image model is installed for ComfyUI.
4. Check available memory, since image models are demanding and commonly need more than language models.
Image models are managed separately from language models. Installing an Ollama model adds nothing to Image Studio.
***
### Still stuck
Record your operating system, browser, selected model, and the exact message shown. If you seek community help, open a current community destination through [Verified Links](/wp/resources/verified-links.md). YveChat does not currently publish a dedicated support or security-reporting channel.
---
# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.
## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.
Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:
```
GET https://yvechat.gitbook.io/wp/getting-started/troubleshooting.md?ask=&goal=
```
`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.
Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.