--- name: opendataloader-pdf-f8458883faf4 description: >- OpenDataLoader PDF converts PDF files into Markdown, JSON with bounding boxes, HTML or plain text, and can add accessibility tags to untagged PDFs. It runs locally on Java with a CLI and Python, Node.js and Java APIs. Use when a user asks to parse PDFs for RAG, convert a PDF to Markdown, extract tables or headings with page coordinates, OCR scanned PDFs, or produce Tagged PDFs. license: Apache-2.0 compatibility: "Java 11+ on PATH, plus Python 3.10+ or Node.js 22.13+" metadata: author: terminal-skills version: "1.1.0" category: data-ai tags: "pdf, parsing, document-processing, rag, data-extraction" repository: https://github.com/opendataloader-project/opendataloader-pdf use-cases: "Extract structured data from invoices, contracts, and reports, Parse PDFs for RAG ingestion with table and image extraction, Build an automated PDF processing pipeline for document analysis" agents: "claude-code, openai-codex, gemini-cli, cursor" --- # OpenDataLoader PDF — AI-Ready Document Parsing ## Overview OpenDataLoader PDF is an open-source (Apache-2.0) PDF parser. A Java engine reads the PDF and writes Markdown, JSON, HTML, plain text, an annotated PDF or a Tagged PDF; the Python and Node.js packages bundle the JAR and start it for you. The default mode is deterministic, runs on CPU and sends nothing to a network. An optional hybrid mode routes hard pages (borderless tables, scans, formulas, charts) to a local Docling-based server. Every JSON element carries its type, page number and bounding box, so an answer in a RAG pipeline can point back to the exact place in the source file. ## Instructions ### Step 1: Install Java 11 or newer must be on `PATH` — the packages do not ship a JVM. Without it the CLI stops with `'java' command not found`. ```bash java -version # must print 11 or higher sudo apt install default-jre-headless # Debian/Ubuntu (Java 11, 17 or 21 depending on the release) brew install --cask temurin # macOS pip install -U opendataloader-pdf # Python 3.10+: library and CLI npm install @opendataloader/pdf # Node.js 22.13+: library and CLI ``` Both packages install the same `opendataloader-pdf` command (`npx opendataloader-pdf` in a Node project). ### Step 2: Convert from the command line Pass any mix of files and folders. The default format is `json`, and output lands next to the input unless `-o` is given. ```bash opendataloader-pdf reports/ invoice-2291.pdf -o parsed -f markdown,json # One file to stdout (single format only) opendataloader-pdf invoice-2291.pdf -f markdown --to-stdout -q # Only some pages, with a page marker in the Markdown opendataloader-pdf handbook.pdf -o parsed -f markdown --pages "1,3,5-7" \ --markdown-page-separator "--- page %page-number% ---" # Encrypted file opendataloader-pdf contract.pdf -o parsed -p "$CONTRACT_PDF_PASSWORD" ``` | Option | Values | Purpose | |--------|--------|---------| | `-f, --format` | `json`, `markdown`, `html`, `text`, `pdf`, `tagged-pdf` | Comma-separated output formats (`pdf` = annotated PDF for visual debugging) | | `--image-output` | `external` (default), `embedded`, `off` | Image files next to the output, Base64 data URIs, or no images | | `--table-method` | `default`, `cluster` | Border-based detection, or border plus text clustering | | `--use-struct-tree` | flag | Read structure from the PDF's own tags instead of guessing | | `--include-header-footer` | flag | Keep page headers and footers (dropped by default) | | `--keep-line-breaks` | flag | Preserve original line breaks | | `--markdown-with-html` | flag | Allow HTML in Markdown for tables with merged cells | | `--sanitize` | flag | Replace emails, phone numbers, IPs, card numbers and URLs with placeholders | | `--threads` | integer | Pages in parallel; values above 1 are experimental | Run `opendataloader-pdf --help` for the full list. ### Step 3: Use the Python or Node.js API `convert()` takes the same options as the CLI in snake_case (Python) or camelCase (Node.js). It writes files: Python returns `None` and Node.js resolves to the CLI's console output, so read the results from the output directory. ```python import opendataloader_pdf # One call for the whole batch: every convert() starts a JVM opendataloader_pdf.convert( input_path=["reports/", "invoice-2291.pdf"], output_dir="parsed/", format="markdown,json", image_output="off", quiet=True, ) ``` ```typescript import { convert } from '@opendataloader/pdf'; await convert(['reports/', 'invoice-2291.pdf'], { outputDir: 'parsed/', format: 'markdown,json', }); ``` For LangChain, `pip install -U langchain-opendataloader-pdf` provides `OpenDataLoaderPDFLoader(file_path=[...], format="text")`. ### Step 4: Read the JSON The top-level object holds document metadata (`file name`, `number of pages`, `title`, `author`) and `kids`, the elements in reading order. Field names contain spaces. ```json { "type": "heading", "id": 42, "page number": 1, "bounding box": [72.0, 700.0, 540.0, 730.0], "heading level": 1, "font": "Helvetica-Bold", "font size": 24.0, "content": "Introduction" } ``` - `type` is `heading`, `paragraph`, `table`, `list`, `image`, `caption` or `formula`; with `--include-header-footer` also `header` and `footer`. - `bounding box` is `[left, bottom, right, top]` in PDF points (72 pt = 1 inch); `page number` starts at 1. - A `table` has `number of rows`, `number of columns` and `rows[].cells[]`; each cell has `row span`, `column span` and its text in `kids[].content`. - A `list` has `list items[]`; an `image` has `source` (the extracted file, absent with `--image-output off`) and `alt_source`. ### Step 5: Hybrid mode for scans, complex tables, formulas and charts Hybrid mode needs a second process: a local server built on Docling. It loads its models at startup, so start it once and keep it running. ```bash pip install -U "opendataloader-pdf[hybrid]" # Terminal 1: the server. It binds 0.0.0.0 by default and has no authentication. opendataloader-pdf-hybrid --host 127.0.0.1 --port 5002 # Terminal 2: the client sends complex pages to it, simple pages stay local opendataloader-pdf --hybrid docling-fast scans/ -o parsed -f markdown,json ``` | Need | Server flag | Client flag | |------|-------------|-------------| | Scanned PDF | `--force-ocr` | — | | Non-English scan | `--force-ocr --ocr-lang "ko,en"` (EasyOCR codes) | — | | Another OCR engine | `--ocr-engine tesseract --ocr-lang "kor,eng"` (Tesseract and its language data must be installed) | — | | Formulas as LaTeX | `--enrich-formula` | `--hybrid-mode full` | | Chart and image descriptions | `--enrich-picture-description` | `--hybrid-mode full` | | Nested heading levels | `--heading-hierarchy` | — | The client looks for the server at `http://localhost:5002`; use `--hybrid-url http://localhost:5003` to point it at another port or host. If the server is unreachable the run fails; add `--hybrid-fallback` to fall back to the local engine instead. ### Step 6: Tagged PDFs ```bash # Add structure tags to untagged PDFs (writes policy-handbook_tagged.pdf) opendataloader-pdf policies/ -o accessible -f tagged-pdf # Trust the tags an author already put in the file opendataloader-pdf annual-report.pdf -o parsed -f markdown --use-struct-tree ``` Auto-tagging produces a Tagged PDF. Export to PDF/UA-1 or PDF/UA-2 is a paid enterprise add-on, not part of the open-source package. ## Examples ### Example 1: Section chunks with citations for a RAG index **User request:** "Parse the PDFs in `papers/` and split them into chunks by section. I need the page and position of every chunk so answers can cite the source." ```bash opendataloader-pdf papers/ -o parsed -f markdown,json --image-output off -q ``` ```python import json from pathlib import Path doc = json.loads(Path("parsed/moran-scene-text.json").read_text()) chunks, current = [], None for el in doc["kids"]: if el["type"] == "heading": current = {"section": el["content"], "page": el["page number"], "text": [], "boxes": []} chunks.append(current) elif el["type"] in ("paragraph", "caption") and current: current["text"].append(el["content"]) current["boxes"].append((el["page number"], el["bounding box"])) for c in chunks[2:6]: words = len(" ".join(c["text"]).split()) print(f'p{c["page"]} {c["section"]} ({words} words, {len(c["boxes"])} boxes)') ``` Output for a 15-page two-column paper: ``` Processed 2 PDF files in 'papers'. p1 Abstract (207 words, 5 boxes) p1 1. Introduction (585 words, 14 boxes) p3 2. Related Work (532 words, 8 boxes) p3 3. Methodology (73 words, 1 boxes) ``` Each entry in `boxes` is `(page, [left, bottom, right, top])`, ready to store as chunk metadata and to highlight in a PDF viewer. ### Example 2: Tables from selected pages as CSV **User request:** "Pull the tables on pages 4 to 6 of `moran-scene-text.pdf` into CSV files." ```bash opendataloader-pdf papers/moran-scene-text.pdf -o tables -f json --pages 4-6 -q ``` ```python import csv import json doc = json.load(open("tables/moran-scene-text.json")) def cell_text(cell): return " ".join(k.get("content", "") for k in cell["kids"]).strip() for el in doc["kids"]: if el["type"] != "table": continue name = f'tables/p{el["page number"]}-table-{el["id"]}.csv' with open(name, "w", newline="") as f: csv.writer(f).writerows([cell_text(c) for c in row["cells"]] for row in el["rows"]) print(name, el["number of rows"], "rows x", el["number of columns"], "columns") ``` ``` tables/p4-table-9.csv 13 rows x 3 columns tables/p6-table-175.csv 16 rows x 3 columns ``` ```csv Type,Configurations,Size Input,-,1×32×100 MaxPooling,"k2, s2",1×16×50 Convolution,"maps:64, k3, s1, p1",64×16×50 ``` The header reads `Configurations` because the PDF stores an "fi" ligature; see the note on ligatures below. If a table comes back split or missing, retry with `--table-method cluster`, then with hybrid mode. ## Guidelines - **Batch the inputs.** Each CLI run or `convert()` call starts a JVM. Pass all files and folders in one call instead of looping over files. - **Start with the default mode.** It is fast and reproducible. Move to hybrid only for scans, borderless or nested tables, formulas and charts; local table detection relies on visible borders. - **Scanned PDFs yield no text without hybrid OCR.** OCR quality depends on the engine, language and scan; try a few sample pages with different `--ocr-engine` values before a large run. - **Keep the hybrid server private.** Start it with `--host 127.0.0.1`; if other machines must reach it, put it behind your own authentication and set `--max-file-size` and `--picture-area-threshold` to bound the work one upload can request. - **`--use-struct-tree` overrides `--hybrid`.** On a tagged PDF the structure tree wins and the backend is not called. Tag quality varies, so compare against the default mode on a sample. - **Content safety filters are on by default.** Hidden text, off-page content and invisible layers are dropped, which protects an LLM from prompt injection hidden in a PDF. Use `--content-safety-off all` (or one filter: `hidden-text`, `off-page`, `tiny`, `hidden-ocg`, `background`) only when you need that content, and treat it as untrusted. - **Text is returned as the PDF encodes it.** Ligatures such as `fi` survive extraction; run `unicodedata.normalize("NFKC", text)` before keyword search or embedding. - **`--sanitize` is a convenience, not a compliance tool.** On 2.5.12 it replaced emails, IPs, URLs and card numbers, but left phone numbers such as `+1 415-555-0134` and a card number written with spaces untouched. Check the output before sharing documents that contain personal data. - **Pass passwords through an environment variable**, never as a literal in a script or shell history. - **Limits:** PDF input only (no Word, Excel or PowerPoint). Heading levels in hybrid mode are flat unless the server runs with `--heading-hierarchy`. Versions before 2.0 were licensed under MPL-2.0. --- Source attribution: terminal-skills Original project: https://github.com/TerminalSkills/skills/blob/8a9852d489f274538be9bf1c9121aeb403750b1c/skills/opendataloader-pdf/SKILL.md License: Apache-2.0 License information: https://www.apache.org/licenses/LICENSE-2.0 Original upstream instructions; downloaded from Mega AI Hub. ## Upstream license notice Apache License Version 2.0, January 2004 http://www.apache.org/licenses/ TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION 1. Definitions. "License" shall mean the terms and conditions for use, reproduction, and distribution as defined by Sections 1 through 9 of this document. "Licensor" shall mean the copyright owner or entity authorized by the copyright owner that is granting the License. "Legal Entity" shall mean the union of the acting entity and all other entities that control, are controlled by, or are under common control with that entity. For the purposes of this definition, "control" means (i) the power, direct or indirect, to cause the direction or management of such entity, whether by contract or otherwise, or (ii) ownership of fifty percent (50%) or more of the outstanding shares, or (iii) beneficial ownership of such entity. "You" (or "Your") shall mean an individual or Legal Entity exercising permissions granted by this License. "Source" form shall mean the preferred form for making modifications, including but not limited to software source code, documentation source, and configuration files. "Object" form shall mean any form resulting from mechanical transformation or translation of a Source form, including but not limited to compiled object code, generated documentation, and conversions to other media types. "Work" shall mean the work of authorship, whether in Source or Object form, made available under the License, as indicated by a copyright notice that is included in or attached to the work (an example is provided in the Appendix below). "Derivative Works" shall mean any work, whether in Source or Object form, that is based on (or derived from) the Work and for which the editorial revisions, annotations, elaborations, or other modifications represent, as a whole, an original work of authorship. For the purposes of this License, Derivative Works shall not include works that remain separable from, or merely link (or bind by name) to the interfaces of, the Work and Derivative Works thereof. "Contribution" shall mean any work of authorship, including the original version of the Work and any modifications or additions to that Work or Derivative Works thereof, that is intentionally submitted to the Licensor for inclusion in the Work by the copyright owner or by an individual or Legal Entity authorized to submit on behalf of the copyright owner. For the purposes of this definition, "submitted" means any form of electronic, verbal, or written communication sent to the Licensor or its representatives, including but not limited to communication on electronic mailing lists, source code control systems, and issue tracking systems that are managed by, or on behalf of, the Licensor for the purpose of discussing and improving the Work, but excluding communication that is conspicuously marked or otherwise designated in writing by the copyright owner as "Not a Contribution." "Contributor" shall mean Licensor and any individual or Legal Entity on behalf of whom a Contribution has been received by the Licensor and subsequently incorporated within the Work. 2. Grant of Copyright License. Subject to the terms and conditions of this License, each Contributor hereby grants to You a perpetual, worldwide, non-exclusive, no-charge, royalty-free, irrevocable copyright license to reproduce, prepare Derivative Works of, publicly display, publicly perform, sublicense, and distribute the Work and such Derivative Works in Source or Object form. 3. Grant of Patent License. Subject to the terms and conditions of this License, each Contributor hereby grants to You a perpetual, worldwide, non-exclusive, no-charge, royalty-free, irrevocable (except as stated in this section) patent license to make, have made, use, offer to sell, sell, import, and otherwise transfer the Work, where such license applies only to those patent claims licensable by such Contributor that are necessarily infringed by their Contribution(s) alone or by combination of their Contribution(s) with the Work to which such Contribution(s) was submitted. If You institute patent litigation against any entity (including a cross-claim or counterclaim in a lawsuit) alleging that the Work or a Contribution incorporated within the Work constitutes direct or contributory patent infringement, then any patent licenses granted to You under this License for that Work shall terminate as of the date such litigation is filed. 4. Redistribution. You may reproduce and distribute copies of the Work or Derivative Works thereof in any medium, with or without modifications, and in Source or Object form, provided that You meet the following conditions: (a) You must give any other recipients of the Work or Derivative Works a copy of this License; and (b) You must cause any modified files to carry prominent notices stating that You changed the files; and (c) You must retain, in the Source form of any Derivative Works that You distribute, all copyright, patent, trademark, and attribution notices from the Source form of the Work, excluding those notices that do not pertain to any part of the Derivative Works; and (d) If the Work includes a "NOTICE" text file as part of its distribution, then any Derivative Works that You distribute must include a readable copy of the attribution notices contained within such NOTICE file, excluding any notices that do not pertain to any part of the Derivative Works, in at least one of the following places: within a NOTICE text file distributed as part of the Derivative Works; within the Source form or documentation, if provided along with the Derivative Works; or, within a display generated by the Derivative Works, if and wherever such third-party notices normally appear. The contents of the NOTICE file are for informational purposes only and do not modify the License. You may add Your own attribution notices within Derivative Works that You distribute, alongside or as an addendum to the NOTICE text from the Work, provided that such additional attribution notices cannot be construed as modifying the License. You may add Your own copyright statement to Your modifications and may provide additional or different license terms and conditions for use, reproduction, or distribution of Your modifications, or for any such Derivative Works as a whole, provided Your use, reproduction, and distribution of the Work otherwise complies with the conditions stated in this License. 5. Submission of Contributions. Unless You explicitly state otherwise, any Contribution intentionally submitted for inclusion in the Work by You to the Licensor shall be under the terms and conditions of this License, without any additional terms or conditions. Notwithstanding the above, nothing herein shall supersede or modify the terms of any separate license agreement you may have executed with Licensor regarding such Contributions. 6. Trademarks. This License does not grant permission to use the trade names, trademarks, service marks, or product names of the Licensor, except as required for reasonable and customary use in describing the origin of the Work and reproducing the content of the NOTICE file. 7. Disclaimer of Warranty. Unless required by applicable law or agreed to in writing, Licensor provides the Work (and each Contributor provides its Contributions) on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied, including, without limitation, any warranties or conditions of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A PARTICULAR PURPOSE. You are solely responsible for determining the appropriateness of using or redistributing the Work and assume any risks associated with Your exercise of permissions under this License. 8. Limitation of Liability. In no event and under no legal theory, whether in tort (including negligence), contract, or otherwise, unless required by applicable law (such as deliberate and grossly negligent acts) or agreed to in writing, shall any Contributor be liable to You for damages, including any direct, indirect, special, incidental, or consequential damages of any character arising as a result of this License or out of the use or inability to use the Work (including but not limited to damages for loss of goodwill, work stoppage, computer failure or malfunction, or any and all other commercial damages or losses), even if such Contributor has been advised of the possibility of such damages. 9. Accepting Warranty or Additional Liability. While redistributing the Work or Derivative Works thereof, You may choose to offer, and charge a fee for, acceptance of support, warranty, indemnity, or other liability obligations and/or rights consistent with this License. However, in accepting such obligations, You may act only on Your own behalf and on Your sole responsibility, not on behalf of any other Contributor, and only if You agree to indemnify, defend, and hold each Contributor harmless for any liability incurred by, or claims asserted against, such Contributor by reason of your accepting any such warranty or additional liability. END OF TERMS AND CONDITIONS Copyright 2025 Terminal Skills Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at http://www.apache.org/licenses/LICENSE-2.0 Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License. # Mega AI Hub resource metadata Creator / publisher: terminal-skills Original project: https://github.com/TerminalSkills/skills/blob/8a9852d489f274538be9bf1c9121aeb403750b1c/skills/opendataloader-pdf/SKILL.md Attribution basis: Source repository record Fields of work: Data & Analytics, Development & Engineering, Product Management Roles: Data Analyst, ML / AI Engineer, Product Manager Specialties: API, Backend Tools: Pdfplumber, Pymupdf, Tesseract, Easyocr, Camelot Country applicability: Country not specified Classification basis: Source category, role and specialization labels Instruction SHA-256: ce6466c39231d32bfc433966379ecf29989c8c22ced9a03838b136faf3ef6a00