Power Up 05: AI Settings Profiles
(AVAILABLE IN ALL EDITIONS OF COTRANSLATORAI)
The "Variable Intelligence" Engine
Stop Using a Ferrari to Deliver Pizza. Match the "Brain" to the Job Instantly.
One of the fastest ways to leak money is using a heavy, expensive AI model for a simple, cheap task. But switching between them has always been a technical nightmare of menus and parameters.
Until now.
We have installed a "Rapid-Switching" Gearbox for your AI. Now, you can define specific "Personalities" for your software—like a "Cheap/Fast" mode for rough drafts, and a "Deep Reasoning" mode for complex problem-solving. You set the parameters once, and then you switch between them quickly.
Watch the briefing to see how to orchestrate your AI resources for maximum profit:
The Slides
We are introducing a major structural change to how CotranslatorAI manages configurations: AI settings profiles.
While the old method of setting models on a prompt-by-prompt basis worked for a while, we realized it created three specific problems that we needed to solve.
First is the "Parameter Explosion." Until now, you only had to worry about model and temperature at the prompt level. But with the rise of "reasoning" models and other new variables, we simply cannot keep adding more knobs and dials to the main window without making the interface unusable. Profiles give us a dedicated space to manage these complex new parameters.
Second is the "Maintenance Nightmare." Previously, if you set a specific model in fifty different prompts in your library, and that model became deprecated or too expensive, you had to manually find and update every single one of those files. By removing these settings from the Prompt Library and centralizing them in profiles, you can now update a setting in one place, and it instantly applies to every workflow using that profile.
Third is "Human-Readable Naming." It is hard to remember the specific use case for "GPT-4o-2024-08-06." It is much easier to remember a profile named "High-Cost Precision" or "Fast Drafting." Profiles allow you to name your settings based on the job, not the tech.
Finally, this architecture gives CotranslatorAI a robust underlying structure for the future. As we look to integrate completely new model families or providers, this profile system ensures the software is ready to handle them seamlessly.
For most users, the standardized default profile will work out of the box. But for those of you pushing the boundaries, this system provides the power you need.
So, how do you actually use this feature?
You manage everything inside the Options window, under the OpenAI API tab.
Here, you can create a new profile and give it a descriptive name, like "Creative Marketing" or "Strict Legal."
Inside that profile, you bundle all your parameters: your model, your temperature, and any advanced settings.
But, while powerful, the various parameters can be confusing. You shouldn't have to study documentation just to know which reasoning effort setting works for which model, or what to choose for "Max. tokens" or "Freq. penalty".
That is why we’ve implemented the "Auto-set" feature.
Inside an AI settings profile, you will see that most parameters default to "Auto-set." This means OpenAI’s server automatically detects the requirements of your chosen model and applies the settings for you that have been set on the back end. It is plug-and-play. You select the model, and the math gets solved.
But for power users who do want to experiment, you can simply uncheck "Auto-set" to unlock full manual control.
Anyway, once your profile is built, it goes into your profile list with whatever name you give it.
Later, you simply select a profile while in a specific tab in the Main Window, which will link the profile to the tab. This is powerful because it allows you to run a "Drafting" profile in tab 1 and a "High-Precision" profile in tab 2, side-by-side.
Finally, you can edit the available models from within the OpenAI API tab in the Options window.
Click on Edit list… to see the list of available models.
The other grayed-out models are set by us at the time of releasing the current version. And we decide on a recommended default model based on considerations of balancing cost, availability, and power.
However, as the user, you can add additional models and change your default model at will. With this approach, new users who do nothing will still get excellent results, but power users can achieve maximum control.
Ready to Apply the Power of CotranslatorAI in Your Work as a Translator?
©2026 Steven S. Bammel & Korean Consulting & Translation Service, Inc. All Rights Reserved.