它结合了版面检测（PP-DocLayoutV3）和 VLM 文本识别（GLM-OCR），在批量 OCR 中实现领先的速度和成本效率，同时内置云存储支持，并可通过 Modal 一键部署。
cheap-ocr was born from the need to OCR a large corpus of PDF documents with complex layouts quickly and at low cost. It converts PDF documents to markdown and json, including documents with complex layouts and tables. It uses pp-doclayout-v3 for layout detection and GLM-OCR to recognize text. We measured throughput at more than 14 pages per second on our benchmark. For other datasets, we reached up to 20 pages per second. It runs on a single L40S GPU (or similar).
Read our blog to learn all about the optimizations that we apply.
We use 100 PDF documents from the European Medicines Agency to benchmark throughput. The documents are typical business documents with a mix of tables, text, and some figures. All metrics are reported end-to-end including download and upload. Warm-up (model loading and initialization) is excluded from our measurements. We compare cheap-ocr against the official GLM-OCR SDK. The benchmarks ran on a single L40S for both solutions.
| Metric | GLM-OCR SDK | cheap-ocr |
|---|---|---|
| Pages/sec | 5.29 | 14.26 |
| Output tokens/sec | 4,316 | 11,801 |
| Page regions/sec | 75 | 205.4 |
| Feature | Description |
|---|---|
| Fast & cheap | Fastest and cheapest VLM-based OCR that we know of. |
| VLM-based OCR | State-of-the-art text recognition that handles difficult text extraction well. |
| Layout detection | Recognizes bounding boxes for headings, paragraphs, images, tables, and more. |
| OCR from cloud storage | Processes PDFs from and writes outputs to S3 buckets, Azure storage containers, or Google Cloud buckets. |
| One-line deployment | Runs the full pipeline on Modal with a single CLI command. |
| Bare-metal deployment | Runs on any machine with an NVIDIA GPU. |
Log into Modal or register a new account. If you are already logged in, skip this step.
uvx --python 3.12 modal login
Then run cheap-ocr directly with uvx; no project files or persistent install are needed. The local command only needs the modal extra. The GPU stack is installed in the remote Modal image.
# run with local files
# startup takes a few minutes because models need to be loaded and CUDA graphs c
OCR with VLMs running on a single GPU with high throughput
根据分类、Topic 和编程语言匹配的相似项目。
Harness Engineering is a methodology for improving coding agent outputs by carefully crafting the environment around them—providing curated context, tools, and executable constraints that encode an organization’s nonfunctional requirements and cumulative lessons.
A 28.9 million parameter language model runs on an $8 ESP32-S3 microcontroller entirely on-device, generating simple stories at about 9.5 tokens per second.
mimic captures traffic from any iOS or web app and automatically generates a Python client library that lets you call the app's API like a regular library.