Your documents,ready to answer.
Upload PDFs, notes, contracts, and manuals. Ask in plain English. Answers cite the exact passage they came from — and invoices, contracts, and résumés can become structured JSON.
Illustrative screenshot: a question about a vendor contract's cancellation terms is answered with a citation to page 6, and invoice fields appear as validated JSON below.
4 formats
PDF · DOCX · MD · TXT
Page-level
citations on grounded answers
Validated
JSON extraction, 4 schemas
Private
no training on your data
The workspace
Ask across everything. Read the source.
A persistent library, hybrid search, and a source panel beside every answer. No copy-pasting between chats.
Illustrative screenshot of the workspace: a question about a vendor contract is answered with a citation, and the matching source passage is shown in a side panel.
Capabilities
Built for documents you can't afford to misread.
Whether it's a contract, a paper, or a 200-page manual, the answer stays tied to the page it came from.
Ask across your whole library
Hybrid semantic and keyword search finds the right passage across every document you've uploaded, not just the one you remember.
See the source behind each claim
Citations link to the exact page and passage. Open one to read the quote in context, right next to the answer.
“…either party may terminate this agreement with 30 days' prior written notice…”
Extract structured data
Turn invoices, contracts, résumés, and key facts into validated JSON you can export and reuse.
{
"vendor": "Northwind",
"total": 12400,
"due": "2026-11-30"
}Private by default
Row-level security, encrypted storage, and deletion that removes your files and their extracted text.
Smart mode for hard questions
Pro routes complex questions to a stronger model and allows longer multi-step research across your documents.
Up to 5 research steps, up from 3.
How it works
Three steps. No new workflow.
If you can upload a file, you can use CorpusBrain.
Upload
Drop in PDFs, notes, contracts, and manuals. Every page is parsed, chunked, and indexed.
Ask
Chat naturally about everything you've saved — in plain language, no query syntax.
Verify
Every claim links back to the exact passage and page, so you can check the source.
Compare
A document workspace, not just a chat.
General assistants can read a file. CorpusBrain keeps a searchable library, cites the exact passage, and turns documents into structured data.
| Feature | ChatGPT | NotebookLM | CorpusBrain |
|---|---|---|---|
| Answers from your own documents | Yes | Yes | Yes |
| Citations linked to the exact passage | Partial | Yes | Yes |
| Persistent multi-document workspace | Partial | Yes | Yes |
| Schema-validated extraction to JSON | Partial | No | Yes |
| No training on your data | Partial | Yes | Yes |
Your documents are not training data.
Isolated storage, row-level security, and deletion that actually deletes. What you upload stays yours.
Pricing
Start free. Upgrade when you need room.
Every plan includes cited answers, structured extraction, private storage, and no training on your data. No credit card required to begin.
Free
For trying it out on a few documents.
- 50 messages per month
- 10 MB storage
- Hybrid semantic + full-text retrieval
- Inline citations
- Structured extraction to JSON
Pro
For people who work with documents every day.
- No monthly message cap
- 1 GB storage
- Smart mode for complex questions
- Deeper multi-step research
Questions
A few things you might be wondering.
Still deciding? The free plan is the easiest way to try it. Upload one document and ask a question. Start free.
What file types can I upload?
PDF, Word (.docx), Markdown, and plain text. Text-based PDFs work best; scanned PDFs are not supported yet.
How is this different from a general AI chatbot?
General chatbots can read files you attach. CorpusBrain keeps a persistent, searchable document library, cites the exact passage behind each answer, and can extract structured data like invoice fields or contract terms.
Is my data private?
Yes. Your documents are stored in your own isolated database rows and object-storage path. We do not use your content to train AI models.
Which AI models do you use?
We use OpenAI models for chat. Pro users can toggle smart mode to route complex questions to a more capable model.
What happens if I hit the free limits?
The free plan includes 50 messages and 10 MB of storage per month. Pro removes the monthly message cap and increases storage to 1 GB. Fair-use rate limits still apply.
Can I delete my account and data?
Yes. Account deletion in Settings removes your profile, documents, chunks, conversations, messages, and subscription record immediately.
Your documents are already worth asking.
Upload your first file and get a cited answer in minutes. Free forever, no credit card required.