crewAI

Very easy and fast way to create AI agents.
You can create multiple agents in one project.
You can give it memory - but if you might need vector DB to store the memory, and embedding model to create the vector. Check here for more details. ```python from crewai import Crew, Agent, Task, Process

short term

my_crew = Crew( agents=[…], tasks=[…], process=Process.sequential, memory=True, verbose=True )

from crewai import Crew, Agent, Task, Process

Assemble your crew with memory capabilities

my_crew = Crew( agents=[…], tasks=[…], process=”Process.sequential”, memory=True, long_term_memory=EnhanceLongTermMemory( storage=LTMSQLiteStorage( db_path=”/my_data_dir/my_crew1/long_term_memory_storage.db” ) ), short_term_memory=EnhanceShortTermMemory( storage=CustomRAGStorage( crew_name=”my_crew”, storage_type=”short_term”, data_dir=”//my_data_dir”, model=embedder[“model”], dimension=embedder[“dimension”], ), ), entity_memory=EnhanceEntityMemory( storage=CustomRAGStorage( crew_name=”my_crew”, storage_type=”entities”, data_dir=”//my_data_dir”, model=embedder[“model”], dimension=embedder[“dimension”], ), ), verbose=True, )

```sh
	# 1. Make virtual env - and start
	virtualenv venv
	sourch venv/bin/activate

	# 2. get crew AI
	pip install crewai
	pip install 'crewai[tools]'

	# 3. create demo
	crewai create crew demo

	# 4. rust is necessary:
	curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh

	# 4.1 check rust
	rustc --version

	# 5. other dependancy install
	# crewai install only works with python up to 3.12 at the moment.
	# You can use pyenv to get the different python version per project
	crewai install

	# 6. Run
	crewai run

Some context:

By using process=Process.sequential, it automatically passes the previous task result to the next task. This can be configured differently, to only send require parameters. The sequence is set by tasks.yml