LLMGraph is a low-to-no-code platform to build, deploy, and host AI workflows. Design on a visual canvas, or just describe what you want in chat, wiring prompts, code, document search, and models into AI assistants. Ship each to a live REST API and an embeddable chat widget in one click. RAG bots, agents, internal AI tools, all with no infra to manage.
No reviews yetBe the first to leave a review for LLMGraph
Hunter
π
Hey Product Hunt π I'm Abaho, solo founder of LLMGraph.
I kept rebuilding the same plumbing for every AI feature β prompt chaining, RAG over documents, calling a model, wrapping it in an API, embedding a chat box. LLMGraph turns that into a visual graph: drag nodes for prompts, code, document search, and models, wire them together, and hit deploy. You instantly get a live REST endpoint + an embeddable chat widget β no infra to manage.
Don't feel like dragging nodes? Just describe what you want in chat and it builds the graph for you.
Every tier has a 14-day free trial. I'd love your feedback β especially on the canvas and the one-click deploy. What would you build with it?
Report
Love the chat-to-canvas approach, that onramp is going to win a lot of people over. One thing I'd really want though: a built-in eval panel where I can run a batch of test inputs through the graph and see side-by-side outputs from different models or prompt versions, with a quick thumbs up/down to track which changes actually move the needle.
Report
Hunter
@esmatztbΒ Thanks for interacting with the product! We've absolutely considered adding the exact kind of testing infra that you described and it's on the product backlog. We just didn't the implementation time to block release without customer input.
Report
Hunter
If you'd like to follow along or reach out, here's everywhere you can find us:
Love the chat-to-canvas approach, that onramp is going to win a lot of people over. One thing I'd really want though: a built-in eval panel where I can run a batch of test inputs through the graph and see side-by-side outputs from different models or prompt versions, with a quick thumbs up/down to track which changes actually move the needle.
@esmatztbΒ Thanks for interacting with the product! We've absolutely considered adding the exact kind of testing infra that you described and it's on the product backlog. We just didn't the implementation time to block release without customer input.
If you'd like to follow along or reach out, here's everywhere you can find us:
X: https://x.com/llmgraph
Bluesky: https://bsky.app/profile/llmgraph.bsky.social
LinkedIn: https://www.linkedin.com/company/llmgraph
Blog: https://llmgraph.substack.com
And if you have questions about the product, grab 30 minutes with me: https://calendar.app.google/2AHX...