Drag and Drop Tools (and their extensions)

[BradKML]

Here are some of the drag-and-drop LLM tooling that might be good as a baseline on what possible tools would be needed

• https://github.com/langgenius/dify
• https://github.com/langflow-ai/langflow
• https://github.com/infiniflow/ragflow
• https://github.com/FlowiseAI/Flowise

Some questions:

1. Are there some integrations that are inherently more valuable than others?
2. What are the major constraints of pure agentic UX vs DAG-like UX?
3. How can be agents more efficient than DAGs when it comes to semi-established work patterns?

[SmartManoj]

How do you compare those with computer.tldraw.com?

[BradKML]

I am partly thinking about how TLDraw is missing a lot of components from these other drag-and-drops:

• Toolkits https://docs.dify.ai/guides/tools https://docs.langflow.org/components-tools https://docs.flowiseai.com/integrations/langchain/tools
• Vector DB + RAGs + Rerranking https://docs.dify.ai/learn-more/extended-reading/retrieval-augment https://docs.langflow.org/components-vector-stores
• Agents (not gonna elaborate too much here)

But at the same time TLDraw is only good because it has the kind of hand-drawn chic (aesthetics not functionally) attached to it. But a lot of these other professional GUI tools can also have the vector DB in the command line. And the way RAGs are used can vary a lot, like document management vs long term memory for web browsing. That is a different issue altogether.