LiteParse

LlamaIndex's open-source, LLM-free document parser: extract spatial text + bounding boxes from PDFs, Office files, and images entirely locally — no cloud, no API keys, no tokens. Rewritten in Rust for v2.0 and shipped to Rust, Node, Python, and the browser (WASM). Source: run-llama/liteparse, https://github.com/run-llama/liteparse, 2026-06-12

LiteParse is the "fast and light" parser from the team behind LlamaParse. It auto-detects a document's format, picks the right parsing strategy, and returns structured text with per-item layout (bounding boxes) — without sending data anywhere or invoking a model. The pitch in @jerryjliu0's launch: "the world's fastest PDF parser, and more accurate than any other open-source, model-free PDF parser (pymupdf, pypdf, markitdown, pdftotext, opendataloader, pymupdf4llm)." Source: X/@jerryjliu0, 2026-05-27

What v2.0 changed

v1 shipped only as a Node/TypeScript package, which capped latency and distribution. v2.0 rewrote the project in Rust with a single core that propagates to every language binding, so LiteParse now runs as a native Rust crate, Node/TS package, Python package, and a WASM build that runs in the browser and on edge runtimes. Source: LlamaIndex blog "Up to 100x Fast Parsing with LiteParse v2.0 and Rust", https://www.llamaindex.ai/blog/liteparse-v2-0-runs-everywhere, 2026-06-12

How it works

• Hybrid extraction: pull structured embedded text directly from the file, fall back to OCR for scanned regions, and route OCR through bundled Tesseract or an HTTP OCR server such as EasyOCR or PaddleOCR. Source: run-llama/liteparse README, 2026-06-30
• Engine: a custom fork/build of PDFium for PDF rendering plus tesseract-rs as the default OCR backend; Rust ownership of the build pipeline is what unlocks the WASM/edge target. Source: LlamaIndex blog, 2026-06-12
• Formats: PDF; Microsoft Office (.docx/.xlsx/.pptx) and OpenDocument (.odt/.ods/.odp) via LibreOffice; images (.png/.jpg/.tiff) via the OCR path. Source: crates/liteparse README, 2026-06-12
• Output: text, Markdown, screenshots, and JSON with bounding boxes, suitable for downstream chunking/RAG or visual page checks. Source: run-llama/liteparse README, 2026-06-30

The launch screenshot benchmark shows LiteParse v2 fastest in each tested row with OCR disabled: 0.002s for a 1-page PDF, 0.041s for 24 pages, 0.123s for 60 pages, and 0.777s for a 457-page PDF. A later Jerry Liu bookmark claims LiteParse is faster and more accurate than open-source model-free parsers on olmOCR0-bench, opendataloader-bench, and ParseBench; the attached local demo shows a browser UI parsing a NASA PDF into text, Markdown, and JSON without leaving the page. The current repo snapshot is Apache-2.0, default branch main at a8288d09cb6bf93c2f7b5257e9634afd26a4566b, 11,279 GitHub stars, and latest 2.4.0 across npm @llamaindex/liteparse, npm @llamaindex/liteparse-wasm, and GitHub tag docker-v2.4.0; the 2026-06-30 PyPI snapshot had liteparse==2.3.0. Source: X/@jerryjliu0 screenshot, 2026-05-27; Source: X/@jerryjliu0 thread/media, 2026-06-18; Source: npm registry packument and git ls-remote, 2026-07-01

Local artifact review: demo thumbnail and preserved video wiki/assets/x-bookmarks/2067679507126124858/video-01.mp4. The self-thread linked accuracy and speed benchmark screenshots, but the enriched X photo fetch was blocked; the durable text claim is captured here with the source tweet and package-version snapshot. Source: raw/x-bookmarks/enriched/2067679507126124858.json, 2026-06-30

Agent skill route

LiteParse now has an upstream Agent Skill in run-llama/llamaparse-agent-skills: npx skills add run-llama/llamaparse-agent-skills --skill liteparse. The skill's main operational point is not "parse everything"; it is parse once to a temporary text file, then search that file with bounded grep/BM25 windows so agents do not repeatedly dump whole PDFs into context. Kevin's wiki imports that skill as skills/engineering/liteparse/ with the bundled scripts/search.py helper. Source: run-llama/liteparse README; run-llama/llamaparse-agent-skills tag liteparse-1.0.1, 2026-06-30

LiteParse vs LlamaParse vs the parser field

LiteParse deliberately stops at fast, local, model-free extraction. For dense tables, multi-column layouts, charts, handwriting, or low-quality scans, LlamaIndex points users at the cloud LlamaParse product, which adds model-driven structure for production pipelines. Source: run-llama/liteparse README, 2026-06-12

Within Kevin's parsing shelf:

• OmniParse — heavier, GPU-friendly multimodal ingestion (Surya OCR + Florence-2 + Whisper, web crawling). Use when the corpus is image/audio/video-heavy.
• MarkItDown - Universal File-to-Markdown Converter — Microsoft's lightweight pure-Python doc→markdown converter; closest in spirit but Python-only and less spatially precise.
• LiteParse — the speed/portability pick: one Rust core, runs in CI, agent workflows, the browser, or edge, with no tokens spent. Feeds RAG pipelines the same way.

Use LiteParse first for local PDF/layout extraction when speed, privacy, or screenshots matter. Use MarkItDown - Universal File-to-Markdown Converter when broad ordinary file-to-Markdown conversion is enough; escalate to MinerU for hard OCR/table/formula PDFs and OmniParse for multimodal/GPU-heavy corpora. Run the three-gate chain (Brin -> skill-auditor -> reputation) before adopting new external packages in production, per AGENTS.md.

---

Timeline

• 2026-05-27 | @jerryjliu0 (LlamaIndex) announced LiteParse v2 — full Rust rewrite, "world's fastest PDF parser," more accurate than model-free parsers, completely open-source and free. 4,076 likes, 4,561 bookmarks. Source: X/@jerryjliu0, 2026-05-27
• 2026-06-12 | Dedicated page created from dev-tools bookmark absorption; researched repo (run-llama/liteparse, Apache-2.0, ~9K stars) + LlamaIndex v2.0 blog. Source: x-bookmark-absorb, 2026-06-12
• 2026-06-30 | One-by-one bookmark artifact review inspected the launch benchmark image, self-thread, OCR reply, upstream repo, npm/PyPI/Git tags, and the upstream Agent Skill. Refreshed routing to make LiteParse the fast local PDF/layout route and imported the skill into skills/engineering/liteparse/. Source: X bookmark artifact audit, 2026-06-30
• 2026-07-01 | Refreshed package/source snapshot: npm @llamaindex/liteparse and @llamaindex/liteparse-wasm advanced to 2.4.0; GitHub tag docker-v2.4.0 resolves to current HEAD a8288d09cb6bf93c2f7b5257e9634afd26a4566b. Source: npm registry packument; git ls-remote, 2026-07-01
• 2026-06-30 | Added the later benchmark/demo bookmark: self-thread claims accuracy and speed tests; local video shows the browser parser producing text, Markdown, and JSON from a PDF. Version snapshot was 2.3.0. Source: X/@jerryjliu0, 2026-06-18; Source: npm/PyPI/GitHub, 2026-06-30