I’m not quite sure if this tag is appropriate, couldn’t find where to put optimization suggestions
1. Regarding gcli2api Configuration Issues
Recently, due to changes in Google’s policies, more people are using CLI proxies, and many are jumping on board without fully understanding, like me… In some gcli2api project tutorials, after completing the preliminary work and returning to the chat platform to configure the API, when choosing the chat completion source as Google AI Studio, which option should be selected in Omate?
2. Some Optimization Suggestions
When creating a character, do not force the selection of an avatar for the character, or provide a few solid color images as initial options for the avatar. Not every character has a preferred matching avatar when being created.
It would be great if the beautification scheme had a one-click import and export function, including background images, bubble colors, bubble transparency, italic and bold text color settings, etc., and even dividers, headers, and footers, to facilitate sharing templates among experts.
Used background images, dividers, and other materials should be saved in history for easy retrieval later. Additionally, it would be nice if the background images could be adjusted for frosted effect besides brightness.
The beautification scheme mentioned in point 2 should also have a save record interface for easy one-click switching, especially useful for day and night mode switching, not to mention that different styles of characters and stories have their own suitable beautifications.
It would be helpful if the mask could separate the name and remarks. Currently, remarks can only be added in parentheses after the mask name, which helps distinguish them in the mask selection interface when starting a new conversation. However, this leads to issues later on, such as when a status bar or other elements need to call the mask name, the user name and remarks in parentheses are combined as the name.
For example: all my mask users are named Xiaomei, and to distinguish them, the mask names can only be written as “Xiaomei (University Version)”, “Xiaomei (Work Version)”. So when I use “Xiaomei (University Version)” for role-playing, if there’s a status bar in the chat, the AI might recognize the status bar name as “‘Xiaomei (University Version)’'s status bar”, which affects the experience.
In the API configuration interface, after copying the API configuration from the top right corner, it would be better to directly jump to the copied version instead of having to go back to the main API configuration page and search for the new copy. It would also be great if the advanced settings below could pull models after entering the URL and key, just like above.
Regarding the issue of duplicate mask and character names, I use letters as substitutes and indicate the names in the mask and character descriptions. For example:
Character Name: SXHb
In the character description:
{
“Character Name” = “Xia Xiaohe”
}
This way, you can set different dialogues for the same character.
In practice, the character name displayed in the Moments is Xia Xiaohe.
Is it only letters and not Chinese? This way, when creating a new conversation, it’s still not very intuitive. When creating a new conversation, you can only see the character names, unless you select them and click the little pencil to view them one by one.
Chinese is also acceptable, just add numbers or letters to differentiate, and state it clearly in the description. Using letters takes up less space, and if the name is too long, it might not display in one line, affecting aesthetics.
In theory, we only provide support for the official APIs of OpenAI/Gemini/Claude.
Due to differences in implementation, we are not aware of how third-party proxies and reverse proxies are implemented. Please consult the vendor or their community directly.
Moreover, due to technical reasons, many proxies and reverse proxies do not strictly adhere to the corresponding specifications. This means their proxies and reverse proxies are incomplete.
We do not provide support for such issues. To confirm whether it’s an OMate issue or a proxy/reverse proxy issue is very simple: directly connect OMate to the official API. If it works, then it’s a proxy/reverse proxy issue.
Some software is compatible with these proxies and reverse proxies largely because they do not perform data validation according to strict official API specifications.
On the other hand, about 50% of API-related feedback is due to sending prohibited content. If you can pass the test when configuring the API, but encounter issues during actual use, and the error message includes the word “Content,” it is usually due to content violations.
About Suggestions
If you don’t quite understand this requirement, you can simply use a part of the background image as a placeholder, which can be edited at any time later.
A global preset will be prepared later, so this part will not be optimized for now.
Please use the Creator Mode - Floating Ball - Common Text function to add it for reuse at any time.
The issue with the status bar name can be corrected directly through prompts or regular expressions.
Regarding the API configuration issue, I will optimize it in the next version.
