🧹 New AI Model, Enhanced Descriptions
Following a recommendation, I decided to switch from using Groq’s API with the Gemma 2 model to Gemini 2.0 Flash Thinking. The transition was seamless thanks to Gemini’s intuitive API design and the availability of a Go SDK, which made the implementation process remarkably straightforward.

🎂 Sending Messages in the past and more!
According to some sources in close contact with Telegram, and what has already been reported by the Russian Telegram Info, A feature is coming that will allow you to send messages in the past, possibly this feature will be used to reorder messages perhaps in a channel in chronological order, we currently have no code on hand, and cannot publish any due to non-disclosure agreements.

We will keep you updated in case the new TL is released, it should be Layer 202

🧙 Improved Descriptions of Layer Changes
I've made a small update that upgrades the AI model from Gemini 2.0 Flash Thinking Experimental to Gemini 2.5 Flash Preview. After testing it, I can confirm it's a noticeable improvement, so from now on, the automatically generated descriptions of newly added methods/constructors will be more accurate and precise.

📰 Just a reminder: This project was originally created to fill a gap left by Telegram, which often delays updating its official documentation sometimes by weeks or even months. Beyond that, it was also designed to help less experienced users easily discover what's new in client updates, and to support MTProto client developers by providing a fully complete schema that doesn’t miss methods which are otherwise available only on specific platforms, such as mobile clients.