Chatbot Using Langflow — Google AI Studio — Composio
Langflow is an open source Python framework with a visual drag and drop interface to build artificial intelligence applications and AI Agents. This framework is a visual interface that built above LangChain
This practical exercise aims to build a simple chatbot application using Google AI Studio to answer user questions, utilizing Composio to connect to the models provided by Google AI Studio.
Command for Langflow Instalation:
1. uv venv langflow_env — python 3.12
2. langflow_env\Scripts\activate
3. uv pip install langflow composio composio-langchain langchain-community beautifulsoup4 arxiv
4. uv pip install -U langchain-google-genai google-generativeai
5. uv pip install -U google-genai google-generativeai langchain-google-genai protobuf
6. uv run langflow run
Step by Step:
- Open Url https://www.langflow.org/templates/use-langflow-to-analyze-code-and-answer-questions/, and click Use for Free
2. Click Download Flow
3. Open Composio website and login using Github Account
4. Move to Platform Options
5. Enter API Key Menu and Create New API Key, then Enter the name for the API key. After that, checked the write all menu and click create. Don’t forget to copy the API Key that appear to use it in the next step
6. Enter Google AI Studio, and click Create New API. After that, enter name for the API key, choose default gemini project, and copy the API Key that appear
7. Open LangFlow and Choose Create First Flow
8. Click LangFlow Logo on the left top corner, and then upload the JSON Template file that had already downloaded
9. Open the Chat with Codebase, and klik Dismiss All
10. Scroll to Github Components, then Enter Composio API Key that already created and copy before. After that, click connect besides play icon
11. Choose Default Auth Mode (Don’t Change to Auth Mode).
12. Click Profiles, then choose Settings. After that, choose Model Providers, and click on the Google Generative AI
13. Enter the API Key that already copy before and click replace (If there is old configuration). After that, don’t forget to choose the model that wan’t to be use
14. Back to the Agent Component, and choose the Language Model that want to be use. In this practice, will use gemini-3–8-flash
Note: If the view is different, try to delete and import the project again
Get Angky Kurniawan’s stories in your inbox
Join Medium for free to get updates from this writer.
15. Click Playground on the right corner to start using the chatbot
16. The view will be same as the image below
17. Try to asking some question, and the agent will give a respond
Note:
1)Try using LangFlow CLI if there is a error, and do the same step as before. After that, Open localhost:7860 in the web browser
2) If the problem doesn’t solve, try the Python version 3.12 or 3.14
