13 - September - 2025

AI Agent - CLI ChatBot

An intelligent conversational agent powered by LangGraph + Groq.

🌟 Features

1. A user can have chat with the Agent
2. Agent can browse the internet
3. Agent can remember the full conversation

🤖 AI Agent | Learning Context

💡 This is not just a chatbot — it's an AI agent:

• LLM ==> is the brain

• Tools ==> is hand (can perform actions like browsing)

• LLM + Tools ==> AI Agent

• Lang-Graph is especially use for building AI Agent

  • The nervous system that coordinates thoughts & actions

• LLM tell us to execute the tool

  • inside tool LLM send our input as an query
  • so tools send the message into LLM and LLM send us that message by formatted way

• Message have 3 type

  • Human Message
  • AI Message
  • Tool Message
    • after tool execute, generated message send back to the AI

📦 Packages

Packages	Usage
bun add @langchain/core	Core utils (for prompts, messages, etc.)
bun add @langchain/langgraph	Manage stateful conversation flow
bun add @langchain/groq	Connects fast with LLMs
bun add @langchain/tavily	Searching tools for LLM

To install dependencies:

This project was created using bun init

bun install

To run:

bun start

🌐 References

• Js LangChain Doc
• Js LangGraph Doc
• Groq API Key
• Tavily API Key
• Tavily Doc

✅ Final Checklist Before Running

• create .env file with proper key = value

Conceptual Data Flow

Example Questions:-

• when was world war 2 happen?
• when was iphone 17 launched?
• what is the weather in Dhaka?
• what is the current stork market info about dhaka?

Example of:- AI Message Response Structure

{
  "messages": [
    {
      "lc": 1,
      "type": "constructor",
      "id": [
        "langchain_core",
        "messages",
        "HumanMessage"
      ],
      "kwargs": {
        "content": "Hi",
        "additional_kwargs": {},
        "response_metadata": {},
        "id": "d5afcde2-f2e2-4e4d-854d-065f0ade6478"
      }
    },
    {
      "lc": 1,
      "type": "constructor",
      "id": [
        "langchain_core",
        "messages",
        "AIMessage"
      ],
      "kwargs": {
        "content": "Hello! How can I help you today?",
        "additional_kwargs": {},
        "tool_calls": [],
        "invalid_tool_calls": [],
        "usage_metadata": {
          "input_tokens": 746,
          "output_tokens": 31,
          "total_tokens": 777
        },
        "response_metadata": {
          "tokenUsage": {
            "completionTokens": 31,
            "promptTokens": 746,
            "totalTokens": 777
          },
          "finish_reason": "stop",
          "id": "chatcmpl-c322d080-0ec6-48a6-a404-8d00d6a9444b",
          "object": "chat.completion",
          "created": 1757941594,
          "model": "openai/gpt-oss-120b",
          "usage": {
            "queue_time": 0.054119138,
            "prompt_tokens": 746,
            "prompt_time": 0.034286912,
            "completion_tokens": 31,
            "completion_time": 0.063442637,
            "total_tokens": 777,
            "total_time": 0.097729549
          },
          "usage_breakdown": null,
          "system_fingerprint": "fp_ed9190d8b7",
          "x_groq": {
            "id": "req_01k56qqyb8et99z76vh5q3jad5"
          },
          "service_tier": "on_demand"
        },
        "id": "10b8f62a-0a6b-47df-a83d-f2a37f079b8e"
      }
    }
  ]
}

Example of:- Tools Message Response Structure

{
  "messages": [
    {
      "lc": 1,
      "type": "constructor",
      "id": [
        "langchain_core",
        "messages",
        "HumanMessage"
      ],
      "kwargs": {
        "content": "what is the weather in Dhaka?",
        "additional_kwargs": {},
        "response_metadata": {},
        "id": "c4b3b403-5146-4a06-937f-71a355b14e97"
      }
    },
    {
      "lc": 1,
      "type": "constructor",
      "id": [
        "langchain_core",
        "messages",
        "AIMessage"
      ],
      "kwargs": {
        "content": "",
        "additional_kwargs": {
          "tool_calls": [
            {
              "id": "fc_bbf363e1-9f68-4344-a88d-a664def5858c",
              "type": "function",
              "function": {
                "name": "tavily_search",
                "arguments": "{\"query\":\"current weather Dhaka\",\"searchDepth\":\"basic\",\"timeRange\":\"day\",\"topic\":\"news\"}"
              }
            }
          ]
        },
        "tool_calls": [
          {
            "name": "tavily_search",
            "args": {
              "query": "current weather Dhaka",
              "searchDepth": "basic",
              "timeRange": "day",
              "topic": "news"
            },
            "type": "tool_call",
            "id": "fc_bbf363e1-9f68-4344-a88d-a664def5858c"
          }
        ],
        "invalid_tool_calls": [],
        "usage_metadata": {
          "input_tokens": 754,
          "output_tokens": 110,
          "total_tokens": 864
        },
        "response_metadata": {
          "tokenUsage": {
            "completionTokens": 110,
            "promptTokens": 754,
            "totalTokens": 864
          },
          "finish_reason": "tool_calls",
          "id": "chatcmpl-0ba1c0d1-f8dd-469e-9666-e75170ffe4d7",
          "object": "chat.completion",
          "created": 1757941640,
          "model": "openai/gpt-oss-120b",
          "usage": {
            "queue_time": 0.046198005,
            "prompt_tokens": 754,
            "prompt_time": 0.035137925,
            "completion_tokens": 110,
            "completion_time": 0.218349241,
            "total_tokens": 864,
            "total_time": 0.253487166
          },
          "usage_breakdown": null,
          "system_fingerprint": "fp_3a688838c3",
          "x_groq": {
            "id": "req_01k56qsb9wffyabttcnephbm4j"
          },
          "service_tier": "on_demand"
        },
        "id": "8689a237-d0d2-4ad4-8741-989f42b253d9"
      }
    }
  ]
}