## Extract clean data from tricky documents thepi.pe is a package that can scrape clean markdown, multimodal media, and structured data from complex documents. It uses vision-language models (VLMs) under the hood for superior output quality, and works out-of-the-box with any LLM, VLM, or vector database. It can extract well-formatted data from a wide range of sources, including PDFs, URLs, Word docs, Powerpoints, Python notebooks, videos, audio, and more. ## Features - Scrape clean markdown, tables, and images from any document - Scrape text, images, video, and audio from any file or URL - Works out-of-the-box with vision-language models, vector databases, and RAG frameworks - AI-native file-type detection, layout analysis, and structured data extraction - Accepts a wide range of sources, including PDFs, URLs, Word docs, Powerpoints, Python notebooks, GitHub repos, videos, audio, and more ## Get started in 5 minutes Thepipe can be installed via the command line: \`\`\`bash pip install thepipe-api \`\`\` The default install only pulls in CPU-friendly dependencies so it is suitable for constrained environments and CI systems. GPU-enabled libraries such as PyTorch and Triton are left as optional extras. ### Optional extras The package exposes a set of extras so you can opt-in to heavier dependencies on demand: | Extra | Installs | When to use it | | -------------------------- | ----------------------------------------- | ----------------------------------------------------- | | \`thepipe-api[audio]\` | \`openai-whisper\` | Local audio/video transcription via Whisper. | | \`thepipe-api[semantic]\` | \`sentence-transformers\` | Semantic chunking with transformer embeddings. | | \`thepipe-api[llama-index]\` | \`llama-index\` | \`Chunk.to_llamaindex()\` conversions. | | \`thepipe-api[gpu]\` | PyTorch + Whisper + Sentence Transformers | Full GPU acceleration with VLM fine-tuning workloads. | If you are targeting CPU-only machines but still need the extras that depend on PyTorch, install the CPU wheels directly from the PyTorch index first and then add the extra. For example: \`\`\`bash pip install torch==2.5.1+cpu torchvision==0.20.1+cpu torchaudio==2.5.1+cpu \ --index-url https://download.pytorch.org/whl/cpu pip install thepipe-api[semantic] \`\`\` If you need full functionality with media-rich sources such as webpages, video, and audio, you can choose to install the following system dependencies: \`\`\`bash apt-get update && apt-get install -y git ffmpeg python -m playwright install --with-deps chromium \`\`\` and use the global installation with pip: \`\`\`bash pip install thepipe-api[all] \`\`\` ### Default setup (OpenAI) By default, thepipe uses the [OpenAI API](https://platform.openai.com/docs/overview), so VLM features will work out-of-the-box provided you pass in an OpenAI client. ### Custom VLM server setup (OpenRouter, OpenLLM, etc.) If you wish to use a local vision-language model or a different cloud provider, you can provide a custom OpenAI client, for example, by setting the base url to \`https://openrouter.ai/api/v1\` for [OpenRouter](https://openrouter.ai/), or \`http://localhost:3000/v1\` for a local server such as [OpenLLM](https://github.com/bentoml/OpenLLM). Note that uou must also pass the api key to your non-OpenAI cloud provider into the OpenAI client. The model name can be changed with the \`model\` parameter. By default, the model will be \`gpt-4o\`. ### Scraping \`\`\`python from thepipe.scraper import scrape_file # scrape text and page images from a PDF chunks = scrape_file(filepath="paper.pdf") \`\`\` For enhanced scraping with a vision-language model, you can pass in an OpenAI-compatible client and a model name. \`\`\`python from openai import OpenAI from thepipe.scraper import scrape_file # create an OpenAI-compatible client client = OpenAI() # scrape clean markdown and page images from a PDF chunks = scrape_file( filepath="paper.pdf", openai_client=client, model="gpt-4o" ) \`\`\` ### Chunking To satisfy token-limit constraints, the following chunking methods are available to split the content into smaller chunks. - \`chunk_by_document\`: Returns one chunk with the entire content of the file. - \`chunk_by_page\`: Returns one chunk for each page (for example: each webpage, PDF page, or PowerPoint slide). - \`chunk_by_length\`: Splits chunks by length. - \`chunk_by_section\`: Splits chunks by markdown section. - \`chunk_by_keyword\`: Splits chunks at keywords. - \`chunk_semantic\` (experimental, requires [sentence-transformers](https://pypi.org/project/sentence-transformers/)): Returns chunks split by spikes in semantic changes, with a configurable threshold. - \`chunk_agentic\` (experimental, requires [OpenAI](https://pypi.org/project/openai/)): Returns chunks split by an LLM agent that attempts to find semantically meaningful sections. For example, \`\`\`python from thepipe.scraper import scrape_file from thepipe.chunker import chunk_by_document, chunk_by_page # optionally, pass in chunking_method # chunk_by_document returns one chunk for the entire document chunks = scrape_file( filepath="paper.pdf", chunking_method=chunk_by_document ) # you can also re-chunk later. # chunk_by_page returns one chunk for each page (for example: each webpage, PDF page, or PowerPoint slide). chunks = chunk_by_page(chunks) \`\`\` ### OpenAI Chat Integration \`\`\`python from openai import OpenAI from thepipe.core import chunks_to_messages # Initialize OpenAI client client = OpenAI() # Use OpenAI-formatted chat messages messages = [\{ "role": "user", "content": [\{ "type": "text", "text": "What is the paper about?" \}] \}] # Simply add the scraped chunks to the messages messages += chunks_to_messages(chunks) # Call LLM response = client.chat.completions.create( model="gpt-4o", messages=messages, ) \`\`\` \`chunks_to_messages\` takes in an optional \`text_only\` parameter to only output text from the source document. This is useful for downstream use with LLMs that lack multimodal capabilities. > ️ **It is important to be mindful of your model's token limit.** > Be sure your prompt is within the token limit of your model. You can use chunking to split your messages into smaller chunks. ### LLamaIndex Integration Install the optional extra and then call \`.to_llamaindex\`: \`\`\`bash pip install thepipe-api[llama-index] \`\`\` After installation, a chunk can be converted to LlamaIndex \`Document\`/\`ImageDocument\` with \`.to_llamaindex\`. Without the extra, a helpful error is raised instead of failing at import time. ### Structured extraction ️ Note that structured extraction is being deprecated and will be removed in future releases. The current implementation is a simple wrapper around OpenAI's chat API, which is not ideal for structured data extraction. We recommend OpenAI's [structured outputs](https://platform.openai.com/docs/guides/structured-outputs?api-mode=chat) for structured data extraction, or using [Trellis AI](https://runtrellis.com/) for automated workflows with structured data. \`\`\`python from thepipe.extract import extract from openai import OpenAI client = OpenAI() schema = \{ "description": "string", "amount_usd": "float" \} results, tokens_used = extract( chunks=chunks, schema=schema, multiple_extractions=True, # extract multiple rows of data per chunk openai_client=client ) \`\`\` ## Running the test suite Install the base requirements plus any extras you rely on, then execute: \`\`\`bash pip install -r requirements.txt python -m unittest discover \`\`\` Tests that depend on optional extras (Whisper, Sentence Transformers, LlamaIndex) or an OpenAI API key are skipped automatically when the corresponding dependency is unavailable. ## Sponsors Please consider supporting thepipe by [becoming a sponsor](mailto:emmett@thepi.pe). Your support helps me maintain and improve the project while helping the open-source community discover your work. Visit [Cal.com](https://cal.com/) for an open-source scheduling tool that helps you book meetings with ease. It's the perfect solution for busy professionals who want to streamline their scheduling process. Looking for enterprise-ready document processing and intelligent automation? Discover how [Trellis AI](https://runtrellis.com/) can streamline your workflows and enhance productivity. ## How it works ️ thepipe uses a combination of computer-vision models and heuristics to scrape clean content from the source and process it for downstream use with [large language models](https://en.wikipedia.org/wiki/Large_language_model), or [vision-language models](https://en.wikipedia.org/wiki/Vision_transformer). You can feed these messages directly into the model, or alternatively you can chunk these messages for downstream storage in a vector database such as ChromaDB, LLamaIndex, or an equivalent RAG framework. ## Supported File Types | Source | Input types | Multimodal | Notes | | ---------------------------- | ------------------------------------------------------------------------------------ | ---------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | Webpage | URLs starting with \`http\`, \`https\`, \`ftp\` | ️ | Scrapes markdown, images, and tables from web pages. AI extraction available by passing an OpenAI client for screenshot analysis | | PDF | \`.pdf\` | ️ | Extracts page markdown and page images. AI extraction available when an OpenAI client is supplied for complex or scanned documents | | Word Document | \`.docx\` | ️ | Extracts text, tables, and images | | PowerPoint | \`.pptx\` | ️ | Extracts text and images from slides | | Video | \`.mp4\`, \`.mov\`, \`.wmv\` | ️ | Uses Whisper for transcription and extracts frames | | Audio | \`.mp3\`, \`.wav\` | ️ | Uses Whisper for transcription | | Jupyter Notebook | \`.ipynb\` | ️ | Extracts markdown, code, outputs, and images | | Spreadsheet | \`.csv\`, \`.xls\`, \`.xlsx\` | | Converts each row to JSON format, including row index for each | | Plaintext | \`.txt\`, \`.md\`, \`.rtf\`, etc | | Simple text extraction | | Image | \`.jpg\`, \`.jpeg\`, \`.png\` | ️ | Uses VLM for OCR in text-only mode | | ZIP File | \`.zip\` | ️ | Extracts and processes contained files | | Directory | any \`path/to/folder\` | ️ | Recursively processes all files in directory. Optionally use \`inclusion_pattern\` to pass regex strings for file inclusion rules. | | YouTube Video (known issues) | YouTube video URLs starting with \`https://youtube.com\` or \`https://www.youtube.com\`. | ️ | Uses pytube for video download and Whisper for transcription. For consistent extraction, you may need to modify your \`pytube\` installation to send a valid user-agent header (see [this issue](https://github.com/pytube/pytube/issues/399)). | | Tweet | URLs starting with \`https://twitter.com\` or \`https://x.com\` | ️ | Uses unofficial API, may break unexpectedly | | GitHub Repository | GitHub repo URLs starting with \`https://github.com\` or \`https://www.github.com\` | ️ | Requires \`GITHUB_TOKEN\` environment variable | ## Configuration & Environment Set these environment variables to control API keys, hosting, and model defaults: \`\`\`bash # If you want longer-term image storage and hosting (saves to ./images and serves via HOST_URL) export HOST_IMAGES=true # GitHub token for scraping private/public repos via \`scrape_url\` export GITHUB_TOKEN=ghp_... # Control scraping defaults export DEFAULT_AI_MODEL=gpt-4o export DEFAULT_EMBEDDING_MODEL=sentence-transformers/all-MiniLM-L6-v2 export FILESIZE_LIMIT_MB=50 # Max duration (in seconds) for audio transcription export MAX_WHISPER_DURATION=600 # Filesize limit for webpages in mb export FILESIZE_LIMIT_MB = 50 # Credientials for scraping repositories export GITHUB_TOKEN=... \`\`\` ## CLI Usage \`thepipe [options]\` ### AI scraping options \`--openai-api-key=KEY\` To enable VLM scraping, pass in your OpenAI API key \`--openai-model=MODEL\` Model to use for scraping (default is \`DEFAULT_AI_MODEL\`, currently \`gpt-4o\`) \`--openai-base-url=URL\` Custom LLM endpoint, for local LLMs or hosted APIs like OpenRouter (default: https://api.openai.com/v1) \`--ai_extraction\` ️ DEPRECATED; will get API key from \`OPENAI_API_KEY\` environment variable ### General scraping options \`--text_only\` Output text only (suppress images) \`--inclusion_pattern=REGEX\` Include only files whose \_full path\* matches REGEX (for dirs/zips) \`--verbose\` Print detailed progress messages ## Contributing This package is quite opinionated in its design and implementation. Some modules are tightly coupled to the overall architecture, while others are designed to be hacked. Before contributing, please create an issue on GitHub to discuss your ideas and how to best implement them. Pull requests that do not follow this process will be closed.