RAG Converter vs Ragas Dataset Converter
Similar domains, different jobs. One prepares source documents; the other prepares evaluation datasets.
What each site does
| RAG Converter | Ragas Dataset Converter | |
|---|---|---|
| Site | ragconverter.com | ragasconverter.com (this site) |
| Job | Turn documents into chunks and embeddings | Convert, validate and check RAG evaluation datasets |
| Input | PDF, Word, Excel, PowerPoint, Markdown, images, audio | RAGAS, TruLens and LangSmith dataset JSON or JSONL |
| Output | JSONL chunks, manifest, embeddings | Another dataset format, field-change report, readiness report |
| Runs locally | Yes, WebAssembly in the browser | Yes, vanilla JavaScript in the browser |
| Pricing | Free tier plus a paid API | Free, no signup |
RAG Converter (ragconverter.com)
RAG Converter is a document ingestion tool. It takes source documents, splits them into token-sized overlapping chunks, and can attach embeddings or an entity graph. Its output feeds a vector database such as Chroma or Qdrant. The free converter runs in the browser through WebAssembly; the paid API handles larger corpora.
Use it when the problem is getting documents into a RAG pipeline.
Ragas Dataset Converter (this site)
This site works one layer later. It takes an evaluation dataset and converts it between RAGAS (legacy, current and modern schema), TruLens and LangSmith. It also reports every field that was renamed or dropped, validates array lengths and contexts types, and marks which RAGAS metrics the dataset can run. Everything happens in the browser tab; no dataset is uploaded.
Use it when the problem is moving or checking an evaluation dataset.
Which one do you need?
- Building a knowledge base or chunking PDFs? Use RAG Converter.
- Migrating an eval set between RAGAS, TruLens and LangSmith? Use Ragas Dataset Converter.
- Checking whether your data can run Faithfulness or Context Recall? Paste it above and read the readiness panel.
- Doing both? They are complementary: ingest documents with one, prepare and check the evaluation set with the other.
Why the names are easy to confuse
“RAG” is the pipeline architecture; “RAGAS” is an evaluation framework. The domains differ by two letters, and both tools run locally in the browser. They are separate projects. This site is independent and is not affiliated with RAG Converter.
Have a RAGAS, TruLens or LangSmith dataset open? Check it now — no upload, no signup.
Open the dataset converterPrimary sources
Field names and metric definitions on this page follow the upstream projects:
- RAG Converter (ragconverter.com) — the tool this page is compared against
- RAGAS documentation — datasets — what a RAGAS evaluation dataset actually is