Messaging System

The goal of this project was to create a messaging system that could receive user messages on behalf of a business, process them in an intelligent manner, and submit a reply that is grounded in the business's details.

General Overview of System

The main architecture for this system is a decentralized network of LangChain agents. A central Router agent analyzes each incoming message and assigns it a classification code based on predefined categories. The message is then passed to a corresponding sub-agent that specializes in that category. Then for the categories whose information about it is dependent upon the particular business, the sub-agent will Analyze the message to further determine what exactly it pertains to, Generates a draft reply as a response to the message, and finally Enforces rules created by the business on the draft reply for the final output message.

Thus, the incoming messages go through a series of 4 steps(when applicable), Route, Analyze, Generator, and Enforce, in order to create a response. Giving rise to the shorthand name for the architecture, RAGE.

A high level overview of the system is shown below:

graph TD;

Input-Message --> Router-Agent;

Router-Agent --> Classification-Code;
  

Classification-Code --> Sub-Agent-1;

Classification-Code --> Sub-Agent-2;

Classification-Code --> Sub-Agent-N;

Sub-Agent-1 --> Output-Message;

Sub-Agent-2 --> Output-Message;

Sub-Agent-N --> Output-Message;

Loading

Where each sub-agent is a set of LangChain agents(the A, G, and E agents) within some function to process the routed incoming message.

Architecture of a Specialized Sub-Agent

After the router classifies the input message(the R step), it's handed to the relevant sub-agent where(if applicable), it goes through the remaining, AGE steps.

A breakdown of each step is listed below:

• The analyze step will process a input message, relevant to the category it exists within, i.e for the service sub-agent the analyze step will process a service related query. The result of this processing is one or multiple classifications codes, that pertain to specific details the message has queried about. These codes are then used in a dictionary lookup to return details that the generator will need in order to draft its reply

• The generator step will process the inputted message along with the details, and generate a draft response to the users query

• Finally the enforce step is there to ensure that the draft reply has followed all critical rules given by the business for the category, such as pricing rules for the service & pricing sub-agent

Note that the sub-agents are not really "agents" at all, and rather just a bunch of Language Models where the outputs are feed into one another, though they are called sub-agents because of how thats what Langchain calls them.

A diagram showing what a message goes through after being routed to a sub-agent is shown below:

graph TD;

Input-Message --> Analyzer;

Analyzer --> Codes;

Codes --> Detail-Lookup;

Detail-Lookup --> Details;

Details --> Generator;

Input-Message --> Generator;

Generator --> Draft-Reply;

Draft-Reply --> Enforcer;

Enforcer --> Final-Reply;

Loading

The path a message takes

The full path that a message goes through from a POST to a POST, is given below

POST /webhook/<channel>  ->  batching  ->  media processing  ->  router  ->  category agents (A G E)  ->  send reply
                                                                                  memory

Where the message batching is essentially just waiting for the user to stop yappping, and media processing just processes the particular type of media sent from the user.

Setup

1. Install the dependencies:

pip install -r requirements.txt

2. Make a .env file inside the MAMS folder

• Use .env.example for how to structure and fill in your actual values
• Set a channel to 0 to turn it off, it still answers its webhook but never runs the agents.

3. Run it:

python3 mams.py

4. Open the tunnel in a second terminal:

ngrok http [PORT]

5. Give each platform its webhook url, using the ngrok address:

Channel	Webhook url
Instagram	https://<your-ngrok>.ngrok-free.app/webhook/instagram
Messenger	https://<your-ngrok>.ngrok-free.app/webhook/messenger
Blooio	https://<your-ngrok>.ngrok-free.app/webhook/blooio

Adding a category

1. Add the new letter and one line describing it to the Categories list in the router system prompt
2. Make a prompts/<category>/prompts.py and prompts/<category>/examples.py, for that category
3. Make a details/<category>_details.json and load it in agents.py with load_details
4. Create the three agents for it in agents.py
5. Write a run_<category>_agent function in agents.py
6. Add one case for the letter that correspond to the new category mams.py switch statement