AI Tab: Telegram Control

From this tab you can control the panel's data generation via Telegram using natural language, thanks to the integration with LLM models via API.

---

Available models

The following models are available in the selector:

Model	Provider	Type
qwen/qwen3-32b	Groq	Free (with per-minute limit)
meta-llama/llama-4-scout-17b-16e-instruct	Groq	Free (with per-minute limit)
llama-3.1-8b-instant	Groq	Free (with per-minute limit)
openai/gpt-oss-20b	Groq	Free (with per-minute limit)
llama-3.3-70b-versatile	Groq	Free (with per-minute limit)
gpt-4o	OpenAI	Paid

For Groq models you can get a free API key at https://console.groq.com

---

Telegram configuration

• Bot token: Telegram bot token. You can get it by talking to @BotFather on Telegram.

• Chat ID: ID of the chat where the bot will send notifications and receive commands. You can get it using @userinfobot.

The Token and Chat ID fields have an eye button to show/hide the value.

---

Model configuration

• Model: Select the LLM model that will process conversations.

• Max tokens: Maximum number of tokens the model can return in each response. Default: 1024.

• API Key: Access key for the model provider. Has an eye button to show/hide the value.

---

Buttons

• Connect / Disconnect: Starts or stops the bot's connection to Telegram. The button text changes based on the current state.

• Load config: Loads a previously saved configuration (token, chat ID, API key, model) from a .json file.

• Save config: Saves the current configuration to a .json file.

• Load chat: Loads a previous conversation history from a .json file.

• Save chat: Saves the current message history to a .json file.

---

Supported commands via Telegram

• Query task status and count
• Add, save and load tasks
• Run and clear all tasks
• Load and save panel configuration
• Query panel folders and settings
• Query MT5 logs, manage files, check system time

Operation results are delivered asynchronously directly in the Telegram chat.