Skip to content
OPQAI.
Sourced intermediate / 🎓 Academic & Research Free tools

Query PDFs with Trieve CLI and LLM Agent

Job to be done: Querying and retrieving information from PDF documents using an LLM agent in the terminal

🇳🇬 Ways to use this in Nigeria

Ideas to get you started, adapt to your situation.

  • Student

    Upload your CSC101 lecture slides PDF and ask 'Summarize the concept of data structures explained in these slides' to prepare for a test.

  • 9-5 employee

    Upload your company's HR policy PDF and ask 'What is the procedure for requesting annual leave?' to quickly find information without sifting through pages.

  • Entrepreneur

    Upload a market research report PDF about the Nigerian e-commerce sector and ask 'What are the top 3 challenges faced by online retailers in Lagos?' to inform your business strategy.

What you’ll get

You will learn how to upload a PDF document to Trieve CLI and then query it using an LLM agent directly from your terminal. This approach uses a search tool, allowing the LLM to iteratively find and reason about information, which is more effective for knowledge retrieval than simply injecting all text into the context.

Tools you need

  • Trieve CLI (freemium): A command-line interface tool for uploading documents and interacting with LLM agents.
  • LLM (freemium): A large language model, such as Gemini, Claude, or ChatGPT, to process your queries.

Steps

  1. Install Trieve CLI: Follow the installation instructions provided in the Trieve CLI GitHub repository. The excerpt mentions it’s available via npm, so you would typically use npm install -g trieve_cli after installing Node.js and npm. You should see confirmation that the trieve command is installed and available in your terminal.
  2. Upload your PDF document: Navigate to the directory containing your PDF file in your terminal. Then, run the upload command, replacing ./document.pdf with the actual path to your file. 
    trieve upload ./document.pdf
    You should see output indicating the document is being processed and uploaded. The tool will process the PDF into smaller chunks for the LLM to reference.
  3. Ask a question about the document: Once the upload is complete, you can ask your LLM agent a question about the document’s content. Use the ask command followed by your query in quotes. 
    trieve ask "What are the key findings?"
    The LLM agent will process your query, use its search tools to find relevant information within the uploaded document chunks, and then provide an answer. Responses stream back with expandable source references.

Original source

This workflow is based on a Show HN post on Hacker News by user skeptrune, introducing Trieve CLI, a terminal-based LLM agent loop with search capabilities for PDFs. The author shared the project to gather feedback on the approach and CLI design.

Notes & variations

  • Free Tier: Trieve CLI offers a free tier for up to 1,000 document chunks, which is suitable for many small documents or research papers.
  • Common Pitfall: Ensure your PDF is text-based and not an image scan. If it’s an image, you’ll need to use Optical Character Recognition (OCR) software first to convert it into a text-searchable format before uploading.
  • Better Results: For more complex queries, try breaking them down into smaller, more specific questions. You can also experiment with different phrasing to see how the agent’s search and reasoning capabilities adapt.

Keep going

More Academic & Research workflows