Guides and engineering notes on PDF to Markdown conversion, document pipelines for LLMs and agents, and token-efficient context.https://marklipi.com/https://marklipi.com/blog/cut-llm-token-costs-document-pipelines/https://marklipi.com/blog/cut-llm-token-costs-document-pipelines/Practical, ranked techniques for reducing LLM spend when your pipeline processes documents: format conversion, caching, routing, context pruning, and the arithmetic to prioritize them.Tue, 28 Jul 2026 00:00:00 GMThttps://marklipi.com/blog/pdf-to-obsidian-markdown-notes/https://marklipi.com/blog/pdf-to-obsidian-markdown-notes/Get a PDF into your Obsidian vault as a real Markdown note — headings, tables and all — instead of an embedded file you can't link, search, or edit. Works for Logseq and any Markdown notes app.Thu, 23 Jul 2026 00:00:00 GMThttps://marklipi.com/blog/convert-pdf-to-markdown-online-free/https://marklipi.com/blog/convert-pdf-to-markdown-online-free/Free online PDF to Markdown conversion in seconds — and the four things to verify before uploading any document to a converter: retention, watermarks, signup walls, and how it handles tables.Tue, 21 Jul 2026 00:00:00 GMThttps://marklipi.com/blog/langchain-pdf-loaders-compared/https://marklipi.com/blog/langchain-pdf-loaders-compared/PyPDFLoader, PyMuPDFLoader, PDFPlumberLoader, UnstructuredPDFLoader — what each LangChain PDF loader actually does, which to pick, and a Markdown-first pattern that chunks better than any of them.Sat, 18 Jul 2026 00:00:00 GMThttps://marklipi.com/blog/parse-pdfs-for-rag-chunking-guide/https://marklipi.com/blog/parse-pdfs-for-rag-chunking-guide/RAG quality is decided at parse time. How to convert PDFs to Markdown, chunk on structure instead of character counts, and avoid the three parsing mistakes that quietly poison retrieval.Thu, 16 Jul 2026 00:00:00 GMThttps://marklipi.com/blog/pdfplumber-vs-pymupdf-vs-pypdf/https://marklipi.com/blog/pdfplumber-vs-pymupdf-vs-pypdf/The three main Python libraries for extracting text from PDFs, compared honestly: speed, table handling, licensing, and when calling an API beats installing any of them.Tue, 14 Jul 2026 00:00:00 GMThttps://marklipi.com/blog/bank-statement-pdf-to-markdown-csv/https://marklipi.com/blog/bank-statement-pdf-to-markdown-csv/Turn a bank statement PDF into clean, structured transactions — Markdown or CSV — in under a minute, without retyping and without uploading it to a sketchy site that keeps your file.Sat, 11 Jul 2026 00:00:00 GMThttps://marklipi.com/blog/extract-tables-from-pdf/https://marklipi.com/blog/extract-tables-from-pdf/Three working ways to extract tables from a PDF — a free online tool, one curl to an API, and Python — plus why extracted tables lose their structure and how to prevent it.Thu, 09 Jul 2026 00:00:00 GMThttps://marklipi.com/blog/convert-pdf-to-markdown-python-node-curl/https://marklipi.com/blog/convert-pdf-to-markdown-python-node-curl/Convert any text-layer PDF to clean Markdown with one HTTP request — no API key, no library install. Working examples in curl, Python, and Node.js.Tue, 07 Jul 2026 00:00:00 GMThttps://marklipi.com/blog/markdown-vs-pdf-llm-token-costs/https://marklipi.com/blog/markdown-vs-pdf-llm-token-costs/Feeding a raw PDF to an LLM costs 60–90% more tokens than converting it to Markdown first. Here's the arithmetic, why it happens, and when the conversion pays for itself.Tue, 07 Jul 2026 00:00:00 GMThttps://marklipi.com/blog/pdf-text-layer-vs-ocr/https://marklipi.com/blog/pdf-text-layer-vs-ocr/If you can select the text in a PDF viewer, the PDF has a text layer and doesn't need OCR. How to check programmatically, why the distinction matters for accuracy and cost, and how to route documents correctly.Tue, 07 Jul 2026 00:00:00 GMT