Multi agent application using google adk
This is multiagent application using custom agent
custom agent is capable of processing the document using chunks based on the document size to avoid token limitation issues. chunks are processed in sequence to avoid token per miniute issues
db agent is capable of generating sql queries from text
front desk agent/root agent is the one that handles the user queries. It is an autonomous agent and decides what to do based on user query.
db agent and custom agent uses tools for fetching the relevant data. Tools will mimic a fetch call and provide fake data, they can be replaced with proper logic to fetch real data
clone the repo
go to the root directory
create .env file and copy the entities from .env_template
you should provide the OPENAI_API_KEY and GEMINI_API_KEY which are mandatory for inferencing the llms
install uv if not available in the system
run
uv venv
uv sync
adk web
happy browsing