Show HN: RAGsplain – What does your RAG model see before it answers?

[fredthedeve]

Are your RAG models hallucinating? Maybe the issue isn't the LLM, but the retrieval phase.

I built RAGsplain (ragsplain.com) to help debug this. You upload documents (PDFs, audio, YouTube links), choose a retrieval method (semantic, keyword, or hybrid), and see the exact chunks of context passed to the model — match scores included.

Turns out: when retrieval is bad, even the best model can’t think straight.

It’s open and free to use. Would love feedback from folks building RAG pipelines.

[datafog]

I like the idea of this. When I tried a couple things it didn't work. Either too long or nothing happened after I clicked the 2nd button. Maybe an example page would be nice.

[fredthedeve]

Thanks a lot for the feedback! Really appreciate you trying it out. For longer videos, transcription can definitely take a bit... I'm working on moving processing to more powerful servers (currently running off my home machine)

For a quick overview on how it works and to get past those initial steps, this demo video of ragsplain might help:

https://youtu.be/BwuA_e7Xn74

[chrisxlucas]

super cool.. it's interesting to see the small advancements/features foundational model providers are starting to build in for interpretability/explainability. this is way more clear UX/UI with the % match and excerpt.