Smarter, more autonomous chatbots (vision, web review, autonomous output)

Current limitations:

• The chatbot only knows what's attached to the board or added as a node in the chat.

• I can only talk to it through text. It can't look at or analyze images, so I have to describe everything in words.

What I'd like added:

1. Image understanding. Let the bot view and analyze images I share, not just read text.

2. Website review. Let the bot visit a website and review or critique it directly.

3. Autonomous agent behavior. I feed it my best examples and reference work, give it insights or research, and it produces a near-final output (80-90% of my end goal) based on its specialty, knowledge base, and the examples it's been trained on. Instead of just answering, it delivers the actual product: images or content.

4. More flexibility and model/agent choice. Give me more control over how each bot behaves and what it can do.

Inspiration to look at: Poe (Poe AI) builds bots along these lines. There are several agent platforms on the market worth reviewing to see what works best, then bringing the strongest ideas into Poppy bots.