DEDUPEDprivacy. Your document contents are sent to the configured LLM provider; use a local OpenAI-compatible endpoint via OPENAI_BASE_URL if they are confidential.
Who it is forA consultant or team with a pile of internal documents.
The problemInternal reports are scattered across files nobody has time to read end to end.
The outcomeA research report grounded in your local documents.
What you need first
- Python 3.12 or later
- An OpenAI API key and a Tavily API key, or a custom OpenAI-compatible endpoint
- Docker if you prefer the docker compose setup
Steps
- Install the app as in the README Installation steps, or the package with
pip install gpt-researcher - Point it at your folder with
export DOC_PATH="./my-docs" - In the web app, choose My Documents from the Report Source dropdown
- With the pip package, pass report_source as local when creating GPTResearcher
Prompt for your agent
Using assafelovic/gpt-researcher, set `export DOC_PATH="./my-docs"` to <my folder> and produce a report on <question> using report_source local.
Caveat. Supported formats per the README are PDF, plain text, CSV, Excel, Markdown, PowerPoint and Word.
Use this in your AI agent
Ask your agent
Using the Deduped MCP, get the setup steps for assafelovic/gpt-researcher and "Write a report from your own documents"
Add Deduped to Claude Code once
claude mcp add --transport http deduped https://dedupedai.com/mcp
Every Monday: the 5 hottest repos and exactly how to use them. Free, one email a week, unsubscribe anytime.
Get them freeShare: X LinkedIn Hacker News Email Copy link
Third-party repos we don't own. AI-written use cases can be wrong. Disclaimer