User Guide for PaLM AI

This guide is aimed at all users of the PaLM AI chatbot on Telegram. It is meant to be a definitive listing and explanation of all the bot’s features. Additionally, it briefly touches upon the concepts behind the bot’s logic and commands to ensure a basic understanding on the part of the user.

Premise

PaLM AI is a free-to-use all-encompassing modern chatbot solution on Telegram. It can be found under the tag @PaLMAI_Bot or the short form link https://t.me/PaLMAI_Bot.

PaLM is developed by the $PALM team. We launched the token as a means of gaining funding and sharing revenue made from the bot to a community of users. If you haven’t, learn more here: https://palmai.tech

How does it work?

First of all, we can start off lightly noting that PaLM chatbot is a risk-less and minimal learning effort requiring bot. The core concepts are input and output — you will give the bot instructions and the bot will forward the instructions to processing and filters, from where the finalized instructions are sent to Google’s PaLM model and then received, processed again for user experience and outputted.

Input to the bot can be separated in two categories:

1. Command-less free logic input
2. Command requiring parameter specifying input

The first of these is defined as any input to which PaLM will react or respond to that does not include a command, characterized by the ‘/’ symbol in front.

An example of command-less free logic input. Note that tagging of the bot that's required in group contexts is not a command and not treated as one in the bot's logic.

Freeform Input

In Telegram, private and public contexts are terms of the API layer used for all Telegram functions to recognize if activity is taking place in private messages or in a group. Each context is assigned an identifier value which can differ depending on specifics of the context, for example the presuffix “-” is added to the identifiers of groups.

PaLM takes advantage of these identifiers much like the Telegram functions do themselves and decides if it should respond only to a tagged message or to all messages within the context, the former being the case in a group and the latter in private messages. The behaviour is otherwise unaltered, though the developer team reserves the use of some commands in private contexts for technical reasons.

Command Input

Commands can be understood as further instructions for the bot to attain a certain “mode” within your discussion context and prepare itself to behave a specific way. An example of this would be using the /chart command, demonstrated below.

The /chart command, released with version 1.1 is meant for technical analysis of token charts.

As a user enters the command, the bot immediately sets to a state where it will expect a certain condition, e.g. in this situation an image to be sent to it. When the image is correctly sent and received, the bot will perform a follow-up action.

If the image should not for some reason be received, the bot will terminate it’s state polling and return to the default state, which is the 0 state, also known as a no-flag state. In a no-flag state, PaLM will simply expect freeform messages to process.

An example of a flag state utilizing command failing it's process and reverting to default state. The further /ca command is ignored because it is processed as the expected input to the /ca command, which waits in a state after the command for the token contract address to be given separately.

However, it is good to keep in mind that not all commands set the bot to a waiting state, in fact it is technically an unnecessary step, only added for user experience reasons. In PaLM, there are many commands functioning without a waiting flag state, one of these being the /list and /help command, which simply trigger instant responses.

Full list of commands as of update 1.1

• /ca — Awaits for input of a valid ERC-20 contract address to analyze the token in a market context
• /chart — Awaits for an image file to take a chart from and analyze using technical chart predictive and reflective methods. Note this only works for images with a chart.
• /help — Shows all current commands and release version
• /gpt — Shortly answers the question ChatGPT vs PaLM
• /track {walletAddress} — Experimental command which tracks all transactions of the wallet address and analyzes them based on available knowledge. Subscribes the user to notifications as the analysis happens real time. Only for private messages as of 1.1
• /remove {walletAddress} — removes a wallet from being tracked. Experimental
• /list — lists all tracked wallet addresses

Conclusion

This guide is not meant to be too long or fatiguing. As such we will end the current iteration of it here. A separate chapter on staking will be added in the near future with a full image tutorial.

We thank you for showing interest in our project and hope you can enjoy your time using PaLM. We also want to invite you to be a part of our community if you already are not — everyone is welcome regardless of investment status, experience, knowledge or any other standing, as part of our mission is to educate.

With respect to the reader,

PaLM Team

Lead Developer