Why uploading papers one at a time does not work
ChatGPT and Claude both accept file uploads, but working with a large collection of documents creates friction. ChatGPT limits you to 10 files per message on Plus plans, with rolling caps on the free tier. Claude accepts files up to 30MB each, but managing 15 separate uploads and making sure the AI connects ideas across all of them is difficult.
The bigger issue is that AI tools treat each upload as a separate item. They do not always synthesize content across documents the way you need for a literature review. When you merge all your sources into one structured file with clear section headers, the AI reads the entire collection as a coherent body of knowledge, not a pile of disconnected uploads.
Practical example: A PhD student working on a systematic review of 20 papers can select the PDF folder in Riflet, add their own annotations file, and export one bundle. Pasting this into Claude gives the AI full visibility across all sources for cross-referencing, gap analysis, and theme extraction.
How to use Riflet for research workflows
Add your paper folder as a source
Point Riflet at your research folder. It extracts readable text from PDFs, DOCX, and over 100 other file types using local processing. No files leave your computer during this step.
Add your annotations and research notes
Include your own Markdown notes, spreadsheet exports of key findings, or summary documents alongside the papers. Riflet combines primary sources and your analysis into one file.
Check token count and export
Riflet shows a token estimate so you know whether your bundle fits within Claude's 200K or ChatGPT's 128K context window. Deselect less important papers if needed, then export.
Ask the AI to synthesize, compare, or critique
Paste the bundle into your preferred AI and ask it to identify themes, contradictions, methodological gaps, or to draft a literature review section. The AI has all the source material in context.
Supported file types for academic work
Riflet reads PDFs, DOCX, plain text, Markdown, CSV, JSON, and code files natively. PDF extraction handles both text-based and most digitally-created PDFs. For scanned documents, you will get the best results by running OCR first with a dedicated tool, then adding the text output to your Riflet bundle.
If your research includes web sources, you can also add URLs directly. Riflet fetches the page content and includes it in the bundle alongside your local files. This is useful for preprints, blog posts, and institutional pages that are not available as PDFs.
Privacy for sensitive research data
All processing happens locally on your machine. Riflet does not upload your PDFs or notes to any server. There is no account, no login, and no cloud component. The only time your content leaves your computer is when you choose to paste or upload the bundle into an AI chat.
This matters for researchers working with pre-publication data, IRB-restricted materials, or proprietary datasets. Your data stays on your machine until you make an explicit decision to share it with a specific AI tool.
Streamline your literature reviews
Download Riflet free and give AI the full scope of your research in a single upload.
Download Riflet free