Verify every claim against your sources. Label every piece before you publish.

Ask any question across your source library and get answers anchored to exact source spans — not generic web filler. Every claim links back to the passage it came from so you can verify it with one click.

Each node is stamped human, AI-assisted, or AI-generated before it leaves your workspace. Readers and crawlers see exactly what they're trusting — more transparent than any major publishing platform.

Scope the fact-checker to your own notes, uploaded PDFs, and research documents — not just the open web. If the answer isn't in your sources, it says so instead of confabulating.

Select exactly which documents ground the retrieval. A legal researcher can lock it to case files; a journalist to their interview transcripts. The AI stays inside the boundary you set.

No tab-switching or copy-paste. Fact-check a claim, get a cited answer, and continue drafting — all inside the same editor where you're writing. The citation lands inline, not in a separate panel.

When your source set changes, a briefing of suggested questions appears automatically — so you discover gaps in your research before you publish, not after.

How does AI fact-checking work in JustJot.ai?

You select the source documents you want to ground on (notes, PDFs, uploaded files), then ask a question or highlight a claim. The AI retrieves the most relevant passages using vector search and generates an answer with inline citations back to the exact spans. It won't invent an answer if the source doesn't support it.

Can I fact-check against my own private sources, not just the web?

Yes — and that's the point. You pick the source set: your own research documents, uploaded PDFs, notes you've written. The retrieval stays inside that boundary. For web-grounded research you can also use the deep-research op, which works from web sources and still attaches citations.

What is the provenance label?

A stamp on every published node that shows whether it was written by a human, co-written with AI assistance (human-directed and edited), or generated primarily by AI. It's visible to readers and machine-readable in the metadata — building reader trust and meeting emerging publisher disclosure norms.

Does the fact-checker work for legal, medical, or financial content?

Yes, for domain-specific content where you have your own source library. A lawyer can ground the fact-checker on case files and precedent documents; a financial analyst on their research reports. The AI answers from those sources with citations — you still review, but the citation trail makes review faster.

What happens if the AI can't find support for a claim?

It says so. When retrieval doesn't find a supporting passage in your source set, the response flags that the claim is unsupported in the current source scope rather than inventing a plausible-sounding answer.

Is there a free plan?

Yes. Start free — ground the fact-checker on your own sources, get cited answers, and label your work before publishing, with no credit card required. Paid tiers unlock heavier usage, larger libraries, and team Spaces.