Unable to access non-serverless model teknium/OpenHermes-2p5-Mistral-7B. Please visit https://api.together.ai/models/teknium/OpenHermes-2p5-Mistral-7B to create and start a new dedicated endpoint for the model.

(For example, here we used the OpenHermes-2p5-Mistral-7B model. However, the error could be any other.)

This means that you have chosen a rarely used model for which you need to pay for a separate server.

Together.ai has two types of models: serverless (standard, via API) and dedicated (for which you need to order servers from them and pay for the time of use). Accordingly, all you need to do to work with TogetherAI is to choose the serverless model. 

Сообщение «Unable to access non-serverless model» в Together.AI — что это значит? появились сначала на ClipAI.

Encoding error. Open the input file in a text editor (such as Notepad), select "Save As" and then select "UTF-8 without BOM" or "UTF-8" encoding.

A similar error can also appear in scripts on Zennoposter or in various programs: if you have question mark symbols “�” appear anywhere, this solution can also help.

To do this, open the text file in Notepad that you specified as input (for example, with the titles of articles):

Click "File" -> "Save As":

And before saving the file, select "UTF-8 without BOM". If you do not have such a parameter (as, for example, in the screenshot below), select “UTF-8”:

Now everything should work fine.

If, however, you get an error:

An error occurred while loading lists. The index was outside the array boundaries.

Then do the above steps with all the files specified in the input settings.

Сообщение «Ошибка кодировки» появились сначала на ClipAI.

This error appears in the log when ClipAI does not receive a response from the AI, or receives an empty response.

To fix this error:

1. Check your Internet connection and its settings. Your IP address may not be suitable for working with this neural network server.
2. Please try again later. If after some time it passes successfully, it means that there were probably temporary problems on the side of the neural network servers (for example, GPT or Gemini).
3. Try changing (usually increasing) the timeout for waiting for a response from the AI in the “Parameters” tab of the input settings.
4. In very rare cases, this problem may occur due to the topic you are trying to write an article on. Try changing it, and if there are any special characters in it, remove them.
5. If you have a mode enabled that sends all input data in the request (for example, "RAW mode" or "Custom template"), while in the "Input data" tab "All content - one request" is selected, and at the same time there is a lot of text in the input (more than 10 thousand characters) - such an error can occur due to exceeding the number of tokens at the input of the neural network ( “context”), or due to the heavy load on the AI servers in this regard.

Сообщение «Произошла ошибка. Получили пустой ответ» появились сначала на ClipAI.

The Gemini has two kinds of basic quota errors.

1. The usual limit on the number of requests per minute/day:

"code": 429, "message": "Resource has been exhausted (eg check quota).", "status": "RESOURCE_EXHAUSTED", "details": null

This limit will be fixed automatically as soon as you stop exceeding it. You can read about the maximum number of requests per minute and per day at corresponding page of Google AI. At the time of the last update of this page, the free version of Gemini has:

In this case, the limits are summed up. That is, total limit for one API key: 75 requests per minute, or 15950 requests per day. However, to do this, your program must be able to automatically switch the model used while running. ClipAI can do it.

The same response can be given if the quota is not used up, but the load on Gemini servers is increased at the moment. This happens both on free and paid accounts.

2. Limit per region:

Quota exceeded for quota metric 'Generate Content API requests per minute'
and limit 'GenerateContent request limit per minute for a region'
of service 'generativelanguage.googleapis.com' for consumer 'project_number:...'

Despite what is stated in the answer, this error usually means that your key is locked. Wherein, judging by user reviews, in this case, you only need to reissue the API key: the account is often not blocked. This usually happens because the program you are using is too aggressive in using different API keys, for example, to generate a single article. This is prohibited by the Gemini rules, which is why it causes blocking.

The following error usually means the same thing:

API key expired. Please renew the API key.

Сообщение Ошибка квоты Google Gemini появились сначала на ClipAI.

You either need to enter the path to the file containing the authorization data:

Or, if you have one website, enter your access data in the fields below:

Also check other input settings for similar errors. If you cannot find the reason, write to us.

And if you have the error "An error occurred while loading tables. This path format is not supported.", then check if the folder path is set correctly in the "Output Data" tab of the input settings.

When the error "An error occurred while loading lists. Path contains invalid characters." check not only the above, but also whether you have the file/folder path set in quotes. There shouldn't be any. For example, instead of:

"C:\Users\Administrator\Documents."

You need to specify the path without quotation marks:

C:\Users\Administrator\Documents\

Also, the file/folder path should not contain unnecessary characters. Therefore, it is best not to enter the path manually or copy it, but select it directly in the input settings.

Сообщение «Произошла ошибка при загрузке списков. Данный формат пути не поддерживается.» появились сначала на ClipAI.

1. Try enabling "Convert to WebP" in the "Processing" tab of the input settings.
2. Increase the “Timeout for waiting for a response to a request from AI” in the “Settings” tab.
3. Check whether access to your site is configured correctly in the “Publishing” tab.
4. Perhaps you do not have access to the neural network used to generate the image? If so, switch the image source used in the “Publish” tab.

Сообщение «Ошибка при генерации картинки, либо во время её публикации на ваш сервер. Продолжаем без картинок.» появились сначала на ClipAI.

1. Remember that you need to indicate the name of the category (heading), not its ID or URL.
2. The category must exist in a standard post type.
3. Try disabling the Litespeed Fix option in the Publish tab.

Сообщение «Указанная категория не найдена. Продолжаем без неё.» появились сначала на ClipAI.

Usually restarting the ClipAI helps (press the “Start” and “+1” button again). However, if this does not help, you can download an additional program manually:

1. Click ­hereto start downloading.

2. Right-click on the line with the script -> “Project Directory”.

3. Open (or create, if it doesn’t exist) the “bin” folder.

4. If the “bin” folder is not empty, delete all its contents.

5. Move the downloaded file to the “bin” folder. Don't change its name. The file name should not contain “(2)” or additional tags, only “ClipAI_Tokeniser” or “ClipAI_Tokeniser.exe”.

Now you can start ClipAI (press the “Start” and “+1” button again).

The error "Unable to run additional program for work. If the error is often repeated - right-click on the line with the script and then delete the file "ClipAI_Tokeniser.exe".

Сообщение «Невозможно загрузить часть программы» появились сначала на ClipAI.

Please note: this is not a ClipAI error. This is the response that your site returns via the REST API.

1. Most often, this error occurs due to incorrect entry of the application login or password. Check if you have configured your WordPress site correctly according to this instructions? Please note that you need to enter not your profile password from /wp-admin/, but the application password, which is exactly what is created in the above instructions. The login is your standard login from /wp-admin/.

2. This usually happens if you created an “application password” from an account that has lowered rights; or entered the username of this account in the “Publishing” tab. For example, he cannot create posts or publish them to all visitors. Check whether posts from this user are published manually through the admin panel (wp-admin). You can also check this by installing a separate plugin for managing roles and role access.

3. The error may be the result of some security plugin installed on your site. For example, it can prevent posts from being published via the WordPress REST API. Check the settings of such plugins and try disabling them.

4. Sometimes this happens if there is an article in the WP Post Recycle Bin with the same title or URL as the post attempt.

Сообщение «Извините, вам не разрешено создавать записи от лица этого пользователя.» появились сначала на ClipAI.

1. Check the stability of your Internet connection.
2. Click the “Settings” button in the upper-left corner of Zennoposter/Zennobox, go to the “Execution” tab, and:
- "Number of HTTP request attempts" - set to "5".
- "Use an alternative method for transmitting HTTP requests" - enable.
- "Use HTTP Connection Pool" - disable.
3. Try running slightly fewer threads. For example, on a non-server version of Windows, such errors can occur on 100+ threads due to the connection limit. This problem may also be due to a limitation on the number of connections on the side of your Internet provider or VPN (if used).

Сообщение Ошибка «Не получилось создать задачу для ClipAI Plus» появились сначала на ClipAI.

1. This model is not allowed for your account. For example, as of the beginning of 2024, GPT-4 was prohibited for accounts where a bank card was not linked to GPT via API. Please note that ChatGPT and GPT according to the API are two completely different and independent services in terms of payment. That is, ChatGPT Plus will not help here.

2. Temporary problems on the AI server side.

3. If you are using Together AI, and selected “Custom model” - check whether you specified the model correctly in the “Custom model” parameter. You need to specify the selected “Model String for API” with Together AI API pages: for example, "meta-llama/Llama-3-70b-chat-hf".

Сообщение «Скорее всего у вас нет доступа к данной модели. Останавливаем шаблон. Подробнее: The model (…) does not exist or you do not have access to it.» появились сначала на ClipAI.

This error may occur if the AI server is temporarily down. For example, the current state of OpenAI servers can be view here.

This can happen when the AI servers are overloaded, or during maintenance on their part. It almost always has nothing to do with ClipAI or your Internet connection. However, it may be related to the parameters you set: for example, if you changed the “Custom API Endpoint” value in the “Options” tab (should be set to https://api.openai.com/v1/chat/completions by default).

The solution is to waituntil the 503 error stops appearing. ClipAI independently waits for the AI servers to recover without interrupting the writing of the article.

Сообщение Как исправить ошибку 503? появились сначала на ClipAI.

Сообщение «Лимит на 200 сообщений в день!» или «Слишком частые запросы к АРІ за день» появились сначала на ClipAI.

1. Open the .htaccess file using the built-in file manager on your hosting or via FTP, and add the line:

SetEnvIf Authorization "(.*)" HTTP_AUTHORIZATION=$1

If that doesn't help, try adding these lines:

RewriteEngine On RewriteCond %{HTTP:Authorization} ^(.*) RewriteRule .* - [E=REMOTE_USER:%{HTTP:Authorization}]

2. Check whether you can switch between Apache and Nginx on your hosting, and whether you can change the PHP version. Try changing these settings.

Сообщение «Ошибка 401» при публикации на WordPress появились сначала на ClipAI.

If it doesn't help - try the following. Open Zennoposter/Zennobox settings:

Go to the “Execution” tab and check the “Use an alternative method for transmitting HTTP requests” checkbox, and increase the “Number of HTTP request attempts” parameter, for example, to the value “5”.

If this doesn't help either — check if you are using any plugin (for example, for security) that blocks the WordPress REST API.

Also check what appears when you navigate to the browser at https://address-of-your.site/wp-json/ . There should be a JSON response (a lot of code). If there is a 404 error, or a redirect to the main page, most likely you disabled the “WordPress REST API” or “wp-json” in the security plugin or in the .htaccess file. It needs to be turned on.

Or, if you are using your own Apache server, enter the command “sudo a2enmod rewrite”.

Also, according to information from users, this error may be affected by the “Easy Table of Contents” plugin.

Сообщение При публикации на WordPress в логе «Пришёл пустой ответ от сайта на WordPress» или другая похожая ошибка; либо статьи не публикуются, но ошибок нет. появились сначала на ClipAI.

This notification occurs if the maximum number of requests per minute to the AI is exceeded. Current information on limits can be found at OpenAI official website (here it is worth looking at the number of RPM (attempts per minute) and RPD (attempts per day)); and for Google Gemini - on the corresponding page. Other services also usually indicate on their website information about account limits.

Accordingly, to prevent this notification from recurring, you usually need to do one of the following:

1. Reduce the number of threads (if you are using multithreading).
2. Top up the account of the AI service you are using.
3. Add an additional API key in the mode of obtaining a key from a file.

Сообщение «Слишком частые запросы к API в минуту. Ждем 20 секунд и пробуем еще раз.» появились сначала на ClipAI.

• Unfortunately I can not…
• Sorry, I can help...
• Sorry, but I can't...
• I'm sorry, but I can't...
• Unfortunately, generating a large amount of text...
• Sorry, I can not…
• I apologize, but I...

And others similar. Sometimes this happens even if the topic is completely allowed by the rules of the neural network.

If one of the above responses is received, ClipAI resends the same request (up to 3-5 times), with slightly different parameters.

However, if you are sure that your article topic does not violate anything and is completely legal, you can significantly reduce the number of repetitions of this notice in one of the following ways:

1. Try using a different model in the “Parameters” tab of the input settings.
2. Reduce the number of characters in the “Processing” tab.
3. Enable “Priority on...” - “Token Saving” in the “Options” tab.

Сообщение «ИИ отказался генерировать статью, пробуем еще раз. Подробный ответ: ‘(…)’ появились сначала на ClipAI.

1. You have an IP address from a country that is not supported by the service you are using. For example, ­here you can find out about the supported countries of OpenAI (GPT), and ­here — Google Gemini.
2. There are problems with the Internet connection on your PC, or the proxy/VPN that you may be using is not working or is not working correctly.
3. You have changed the "Custom API Endpoint" option in the "Options" tab. In this case, the default value should be returned:

   https://api.openai.com/v1/chat/completions

4. Temporary technical problems on the AI server side.

{ "error": { "code": 400, "message": "User location is not supported for the API use without a billing account linked.", "status": "FAILED_PRECONDITION" } }

{
{ "error": {
{ "code": 400,
"message": "user location is not supported for the api use.",
"status": "failed_precondition"
}
}

Сообщение «Ошибка при выполнении запроса. Проверьте интернет соединение. (…)» появились сначала на ClipAI.

To reduce the number of repetitions of this message in the log, you must do one of the following:

1. Change the operating mode to “Writing an article (with sections, for large articles).” It is recommended to do exactly this: in the mode of writing large articles, completely different, improved algorithms and methods for correcting logical repetitions are used.
2. If you need no more than ~2000 characters in an article, disable the “Continue article if there are not enough characters in one answer” option in the “Processing” tab. In this case, there will be no repetitions when asking to continue the text, since there will be no requests to continue the text.
3. Try increasing "Match check aggressiveness (0-100)". Please note: this is not a percentage of repetitions, but “points”. The number of “points” is determined by the AI after comparing two elements of the article (previous and current).
4. Disable the "Check for matches when continuing text" option.

Сообщение «Обнаружено частое повторение тем: X пунктов. Задано Y пунктов.». Что делать? появились сначала на ClipAI.