top of page
Search
Writer's pictureTim Kosmala

ChatGPT/GPT-4 for WhatsApp, Instagram, Telegram and others…

AutoResponder is the easiest way to combine the state-of-the-art AI ChatGPT, GPT-3 or GPT-4 by OpenAI with WhatsApp, Telegram, Facebook Messenger, Instagram, Viber or Signal. You can have your chats answered automatically with an artificial intelligence that has been trained on most of the internet text and thus offers answers to any query you can imagine. As a chatbot, OpenAI's GPT can perform various tasks including replying to questions, generating text, and much more.


It is known for its flexibility and customization capabilities, allowing you to specify the tone, style, and formality of its replies. You can even tell ChatGPT / GPT-4 whose personality it should assume when replying (e.g. that of a cell phone salesman). Or you define what it should pay special attention to when answering. It's like your personal assistant. Look at the examples and be amazed.


AutoResponder also offers many configuration options, including AI replies to selected messages only and user-defined rules for others. You can get the app for free and without an account. What are you waiting for?


Overview


Integrate ChatGPT / GPT-3 in WhatsApp, Telegram, Facebook Messenger, Instagram, Viber or Signal.

Always note: GPT can sometimes generate wrong information, harmful instructions, or biased content with its limited knowledge.


Spanish? Entonces mira este video de YouTube.


Follow these instructions to set up ChatGPT/GPT-4 and AutoResponder


1. Get AutoResponder. Check the how-to if needed.

2. Create an OpenAI API account. Make sure you have enough funds in your account available here (You need to buy credits through your credit card in advance. Don't worry, it's very cheap and you can start with only a few credits if you like. With the default settings, you can send around 25.000 AI messages per 1 USD).

3. Get your API key from here and copy it.

4. Create a new AutoResponder rule.

5. Tap the All-button at the top right corner if you want the AI to reply to any incoming messages. You can still limit replies to specific contacts. Otherwise, you can use the other Received message features of AutoResponder. Check the info buttons next to each option.

Find the all-button at the top right corner of your AutoResponder rule.

6. Make sure the Reply message field is blank. Then activate the Connect OpenAI's ChatGPT checkbox below.

The Connect OpenAI's ChatGPT/GPT-3 checkbox can be found below the reply message and multiple replies field.

7. The prompt is automatically filled in, based on which the AI will generate an answer later. You can always customize it to suit your needs. For example, add general information about your business that it should refer to when responding. Check here for how to design your prompt and combine it with the answer replacements of AutoResponder (previous messages, name, date, time etc.). But test with the ready-made text for now.

The following is a conversation with an AI assistant using AutoResponder.ai. The assistant is helpful, creative, clever, and very friendly.

Note that chat models like gpt-4o-mini (which is the default) or gpt-4o are chat-completion models and can use special chat prompt formatting.


In AutoResponder, any text before the first 1:: or 2:: is a system message that you can use to tell ChatGPT how to behave and what information to include in the reply.


Use 1:: before any text you want to send with user role (as a user message that has received or should receive an automatic answer) and 2:: for assistant role (the reply message of ChatGPT). These messages in combination with AutoResponders answer replacements are used to send the chat history to the AI along with every request, because it does NOT keep a history by itself. The last %message_512% answer replacement is especially important to tell ChatGPT what message you have received from your contact so that it can reply to that.


If you don't use 1:: or 2::, the prompt will be sent to the AI as a full user message. If you also leave %message_512% away, the prompt will be static. This is useful if you have a rule that for example replies to a keyword like "!joke" and your prompt is only "Tell me a joke".


If you want to use another model like text-davinci--003, which is a text-completion model, you can keep the formatting with 1:: and 2:: to get similar results, but the prompt is always sent as a whole to OpenAI.


8. Now paste the API key you copied earlier below. The other parameters are already pre-filled. You don't have to change them, but you can see how it works here.


Change the model parameter to the version of GPT you want to use (e.g. gpt-4o). Note that OpenAI may not have granted you access to specific models yet. You can also use OpenAI Assistants here (AutoResponder v3.8.0+). Note that Assistants ignore the Stop sequences, presence_penalty and frequence_penalty parameters.


If you do not want the answers to be cut off at a certain length, increase the max_tokens value or tell ChatGPT that it should write shorter replies.


Reply prefix, Policy violation reply and Error reply are custom fields of AutoResponder. For example, you can use the robot emoji 🤖 as an optional reply prefix to show your chat partner that the answer comes from an AI.

All of the custom parameters for your AutoResponder and GPT integration.

9. Save the rule. That's it! You can now test your ChatGPT / GPT-4 and messenger integration and impress your friends or clients 🎉


Tips and Tricks


1. If you are an advanced user, you could even train the AI using your own chat history or custom data. Assistants (AutoResponder v3.8.0+) and GPT fine-tuning are supported by AutoResponder.


2. Improve the reliability of the AI by checking here.


3. If you are using the gpt-4o-mini or gpt-4o model it may sometimes be better to append ChatGPT commands to the last user message to get more consistent results:

Here, "Reply to the following message like a pirate:" was added to the last user message in the prompt.

4. You can use AutoResponder's answer replacements to give the AI more useful information. For example, it can address the user by name if you are using %name% in the prompt. The same is also possible with the date and time.

Add "You write with %name%." to the beginning of the prompt so the AI can address a user by name.

5. OpenAI bills for used tokens. See their pricing. Then you can check here how much tokens a specific text consists of. Note that all tokens of the prompt AND the reply are counted. To save reply tokens you can do several things:

  • Keep your prompt short (for example, by including less of your users' message history).

  • Use a less capable AI model instead of the default gpt-4o-mini (which results in answers that are not as good, but it can be useful for simpler tasks).

  • Reduce the max_tokens parameter (makes the answers shorter).

  • Only use ChatGPT replies for specific received messages.

  • Write a specific prompt for a specific rule without including the incoming message of the user.

  • Reduce the maxlength number of the AutoResponder answer replacements like %message_512%. The "512" can be changed to limit the number of characters.

  • Just create normal AutoResponder rules for messages you receive very often.

  • Be creative :)

With the default settings, you can send around 25.000 AI messages per 1 USD. I think that's already a fair price considering how long it would take to write so many WhatsApp or other messenger messages manually.


Troubleshooting


The troubleshooting moved. Please check here if you face any issues with ChatGPT.

90,396 views

Related Posts

See All
bottom of page