We doesn't let our customers to choose our AI Model, because we standardize  and customize our AI model for cover letter generation, that make this will works in your favor.

We regularly test different advanced models using real job descriptions and carefully designed checks. This process makes sure the model we choose:

After testing, we select the model that performs best and tune it further to improve accuracy and consistency.

By keeping all customers on the same optimized model, we ensure a consistent, high-quality experience. Allowing individual model switching would create uneven results and reduce the overall accuracy of your cover letters.

We don’t expect you to just take our word for it, here’s what the data shows:

Our model scores 0.895 on ROC AUC analysis, that we use as a way of checking how well it tells stronger job matches from weaker ones. Think of it like this: if you put two jobs side by side, the system will pick the better match almost 9 times out of 10. That means your cover letter starts from opportunities that truly fit. We also tune the system for high recall, meaning it’s better to show a few extra jobs that aren’t perfect than risk missing the one that could lead to your next contract.

The F1 score rose from about 0.50 to 0.75–0.78. The system is now much better at finding the right jobs while avoiding irrelevant ones. That’s a 45–50% improvement, making job filtering more reliable.

Customer satisfaction with job filtering has risen from the low 60s% to the mid-70s%. Even though the chart is about filtering, many users give feedback with the cover letters in mind. That makes this improvement a good sign that the cover letters themselves are now clearer, more relevant, and more polished.

We don’t let customers switch models because that could hurt accuracy and lower the quality of your cover letters. Instead, we regularly test, improve, and fine-tune the best model available so you get steady improvements and consistent results.