Integrations and settings

AI Integration

PLEASE NOTE!

AI processing is only possible when delivering messages using the copy method. When delivering messages using the built-in Telegram forward mechanism, no content transformation is possible, so processing in forwarding mode is not available!

The AI allows you to describe in plain language the action you want to perform on the text from a Telegram message from the source. Junction Bot then sends the message and your instructions to the AI, and publishes the processed results to the destination.

Popular use cases include:

translation into another language,
analysis and summarization of long texts,
smart filtering,
changing numbers in the text according to specified criteria,
and much more.

Important: AI processing is applied only to text, image processing is currently unavailable.

‍

You can configure AI processing for each forwarding individually. To do this, go to the settings for the desired forwarding and select the AI processing section:

Where to find it

Menu → My Forwards → Select a Forwarding → Forwarding Settings → AI Processing (GPT and others)

‍

How to set up AI message processing

Guide How to set up AI integration

The main AI processing screen is divided into three logical blocks for convenient management of your AI instructions (prompts).

Active Prompt : This is the prompt that is currently running for this forwarding. If no prompt is applied, this section is hidden.
My Prompts : A list of all the prompts you've created or modified yourself.
Prompt Library : A collection of ready-made templates for popular tasks.

How to use prompts

Using Prompt

To activate a prompt, click on it in the list and then click the “Apply ” button that appears at the bottom of the screen.

After this, the prompt will move to the "Active Prompt" block and begin its work.

Important: You can use the same prompt for multiple forwards. However, if you edit such a prompt, the changes will be applied to all forwards where it is used.

‍

Editing and other actions

Each prompt has a context menu (the icon with three dots). You can use it to Edit , Copy, or Stop the active prompt. Stopping the prompt deactivates it and returns it to its list.

Important: When you edit a prompt from the Library and then save it, a copy of it is always created in the "My Prompts" list. The original template remains unchanged.

‍

How to create your own prompt

If you need a unique instruction for the AI, you can create your own prompt.

In the "My Prompts" block, click + Create Prompt .
The editing screen will open. Fill in the fields. For example, if you want hashtags automatically added to technology news:
- Name: _{Tech hashtags}
- Description: _{Adds #ai, #startup, #crypto to posts about technology.}
- Prompt text: _{Analyze the text. If it's about technology, artificial intelligence, startups, or cryptocurrencies, add relevant hashtags to the end: #ai, #startup, #crypto.}
- Creativity: _0.7

(Creativity (or "temperature") determines the style of the answer: the higher the creativity, the more artistic the answer will be; the lower it is, the more "scientific" and precise the answer will be. 0.7 is recommended).

At the end you have two options:
- Save and Apply : The prompt will be created and immediately activated for the current forwarding, appearing in the "Active Prompt" block.
- Save : The prompt will simply be added to your "My Prompts" list, allowing you to apply it later.

‍

Payment for AI processing

We offer several options for using AI processing so you can choose the one that best suits your needs.

Important: This choice will apply to all your forwarding and other AI services in the bot.

1. Free mode

To help you get familiar with the service, we offer 10 free AI processings. This limit is renewed monthly. In free mode, one default model is used for processing.

‍

2. Payment from balance

This mode provides access to over 300 AI models and uses OpenRouter's provider selection optimization system. It automatically selects the best provider based on cost and speed for your selected model.

Cost: The cost of the call is displayed when you select a model. Funds are deducted from your internal balance in the bot.

‍

3. Use your API key

This mode allows you to connect your own API key from one of the providers. Payments in this case are processed directly through your provider; Junction Bot does not charge a processing fee.

^OpenAI:Your request will be processed through OpenAI, so only models provided by this provider are supported. To connect, please enter your OpenAI API key in the settings.
- Cost: Funds will be charged to your OpenAI account.
^OpenRouter:Your request will be processed through the OpenRouter LLM aggregator, which supports over 300 models from various providers. This list also includes free models, so using your OpenRouter account, you can process messages for free. To use, you must enter your OpenRouter API key in the settings.
- Cost: Funds will be debited from your OpenRouter account.
- Important: Connection via the OpenRouter API key is only available on the Expert plan.

To switch the payment mode, click the mode selection button at the top of the settings page.

‍

Other Bot Settings

Bot Interface Language

‍
To change the bot's interface language, use the _/language command.
Available languages: English and Russian.

Importing and Exporting Settings

‍
Using the _/dump command, you can export all current subscriptions, redirects, and their associated settings as a file. The file will be a list of commands for the bot.

You can execute some or all of these commands at once using the _/batch command. To do this, send the bot a list of commands starting with _/batch, with each line containing a separate command.

Note: When you send a very long text in Telegram, it may be split into multiple messages. In such cases, the bot will only process the commands in the first message (the one that starts with _/batch). To process a very long list of commands, you’ll need to send it in multiple messages yourself, making sure that each of these messages starts with the _/batchcommand.

‍

Integrations and settings

Read our blog

3.9.2025

Privacy and Security in Junction Bot: How Account Connection Works

Is it safe to connect your Telegram account to the bot? 🤔 In our new article, we answer the most frequently asked questions about privacy in Junction Bot. We tell you how your data is protected, why the risk of blocking is minimal, and how to make sure we are reliable.

Account Connection

14.10.2025

How to increase message delivery speed?

Is your Telegram bot slow? Learn why Telegram's priority system causes message delays and discover 3 proven solutions. From using Telegram Premium to creating a dedicated "clean" account, our guide shows you how to increase delivery speed instantly.

16.6.2025

Creating a Self-Updating AI Expert Based on a Telegram Channel

Build an AI expert that automatically updates from Telegram channels using Junction Bot and ChatGPT. Create smart knowledge bases from live feeds, automate learning systems.

Read blog