Experience with MP have been amazing. I don't think I am using the full potential of it yet.
What I think is great:
1. It allows me to create many different agents. Each of them is like a custom GPT. More importantly, for each agent, I can choose the models I want to use.
2. Multiple integrated LLM. I can use OpenAI, Anthropic, Google, Groq.
3. I can create workflows like a process using different agents. It's like a production line! A mixture of different agents doing different work, it is so fun!
4. There are pre-use templates available!
5. Allow workflow builds with API and Webhooks with external tools!
6. Allow sharing workflow / agent as an external tool.
Area of improvement:
1. Transcription model is limited to NOVA V2, will be happy to have multiple transcription model for selection.
2. Javascript is not integrated.
3. There is no specific library or template for Image generation even thought it can.
