@liteai_free_bot — an AI Telegram bot

Today we launched a separate Telegram bot for chatting with language models — @liteai_free_bot. It doesn't replace the main @liteaitech_bot; they're partners: liteaitech_bot handles token purchases and API keys, while liteai_free_bot handles the actual conversation with the model right inside Telegram.

Two modes of operation

The bot runs in one of two modes, and switches between them transparently:

Free mode — the bot's shared key. No registration needed: open the bot, tap Start, and start typing right away. Requests go through a shared LiteAI key. A fair daily request limit applies — to make sure resources are available for everyone. The bot shows the exact limit on first launch. All the ways to get free access to the Claude API (free quotas, LiteAI $1/1M) and their limits — in the article “Trying the Anthropic API for Free in 2026”.

With your own LiteAI key. If you already have an API key from liteai.tech (bought via the main bot or picked up a package on the site), connect it with the /apikey command and the bot switches to "advanced" mode:

• every request (chat and voice) goes through your key, not the shared one;
• the list of available models is fetched directly from the liteai.tech server — the same 9 models you see on /pricing: Claude Opus 4.8 / Sonnet 4.6 / Haiku 4.5 / fable-5 and / / 5.6-luna / 5.6-sol / 5.6-terra;
• the daily request limit is lifted — the only cap is your key balance;
• the bot immediately deletes the message containing your key from the chat, so it never stays in your history.

Switching between modes is free: send /apikey again and the bot goes back to the shared key.

What the bot can do

Chat with conversation memory. The bot keeps full conversational context — every next message is aware of the previous ones. Context length depends on the chosen model — for Opus 4.8 and Sonnet 4.6 it's 200K tokens, enough for long discussions and working with large texts. Reset the context and start over with /reset.

Voice messages. Send the bot a voice message — it transcribes the speech and replies with text. Handy when typing would take too long but you want to capture an idea right now. Transcription works in both modes — shared key and your own.

Model selection. The /model command opens a menu with the available models. If the selected model is temporarily unavailable (provider overload or maintenance), the bot automatically switches to the next working model and tells you about it in the chat. The conversation doesn't drop.

Support from the bot. The /support command opens a chat with the LiteAI admin right inside the bot — usually answered within an hour during business hours. The same channel covers questions about the main liteaitech_bot, key balance, and any technical bits.

How to try it

Open @liteai_free_bot in Telegram and tap Start — that's all you need to start typing in free mode. If you already have a LiteAI API key (sk-bf-…), send it to the bot via /apikey and switch to your model and your balance.

Free mode is a great way to try LiteAI in action before buying tokens: see how the models respond, which conversation style suits you, whether you need voice. And when you decide to move to your own key — /apikey connects it in seconds, with no re-registration and no lost conversation history.

Buying your own key (all the methods, the pricing, the pitfalls) — in the article “How to Buy the Claude API from Russia”. And if you prefer a web chat in the browser instead of Telegram — LiteAI Chat activates automatically on the same purchase.