Use case
Chat with your documents. Verify every answer.
Upload PDFs, Office files, Markdown, or a ZIP of your whole library and chat with your documents—every answer cites the exact source so you can verify it.
How it works
From folder to first answer
Upload the documents
Drop one supported file or a ZIP carrying up to 10,000 documents. Uploads go straight to private storage and resume if the connection drops.
Let the pipeline prepare them
kbDrop checks each file, extracts the text, splits it into passages, and indexes it into one private knowledge base—no retrieval pipeline to set up on your side.
Ask in plain language
Open chat and ask the way you would ask a colleague. Answers stream in with citations to the exact documents they came from.
Formats
Bring the files you already have
kbDrop reads PDF, DOCX, PPTX, XLSX, HTML, Markdown, TXT, and CSV documents, plus supported PNG, JPEG, WAV, and MP4 media. There is no export-to-our-format step: the library you maintain today is the library you chat with.
For a whole folder, pack it into a ZIP. One archive can carry up to 10,000 files, and kbDrop validates and indexes each entry into the same private knowledge base, so a question searches the entire collection at once.
Why citations matter
An answer you cannot check is a guess
Chat-with-your-documents tools live or die on trust. kbDrop attaches citations to every answer, in the browser and over the knowledge base API, so the person reading the answer can open the underlying source and confirm it says what the answer claims.
Your documents stay private to your account, and the security overview spells out exactly what crosses to the model provider and why.
Questions, answered
Chatting with documents FAQ
Can I chat with a PDF?
Yes. PDF is one of kbDrop's core formats, alongside DOCX, PPTX, XLSX, HTML, Markdown, TXT, CSV, and supported PNG, JPEG, WAV, and MP4 media. Upload the file, wait for the knowledge base to become ready, and start asking questions.
How many documents can I chat with at once?
A single ZIP can carry a document library of up to 10,000 files, and everything in it is indexed into one knowledge base. You ask one question and retrieval finds the relevant passages across the whole library.
Do answers show where they came from?
Every answer carries citations pointing at the indexed source material. Instead of trusting a fluent paragraph, a reader can open the cited document and check the evidence.
Will my documents be used to train an AI model?
kbDrop does not use your documents to train its own models. Selected content does reach kbDrop's model provider to index sources, analyze supported media, and generate answers—the security overview documents that boundary so you can choose source material with it in mind.
Do I need to set up a vector database?
No. Validation, extraction, passage splitting, indexing, retrieval, and citation assembly run as one managed pipeline. You upload documents and ask questions; kbDrop operates the retrieval machinery.
Ask your library a question today
The Free plan includes enough source processing and cited answers to test kbDrop on a real document set—no card required.
Create a free account