--- title: "🧪 GPU-Free and 15x Faster: 1.25M Parameter indra-ocr is a Local Savior | Local AI Lab" description: "This issue features indra-ocr, which has only 1.25M parameters and enables high-speed Traditional Chinese OCR on CPU servers completely GPU-free, and the Kromoss image generation model, which quickly generates clear Chinese characters in ComfyUI in just 8 steps. Are you a backend engineer wanting to protect customer privacy without bearing expensive cloud API costs? We will show you how to build a high-throughput document processing workflow locally using the most resource-efficient architecture." canonical_url: "https://blog.markkulab.net/en/local-ai-lab/local-ai-lab-2026-10-21" author: "Mark Ku" author_url: "https://blog.markkulab.net/en/author/mark-ku" site: "Mark Ku's Tech Notes" date_published: "2026-10-21T02:00:00.000Z" category: "On-Premise AI Lab" tags: ["local-ai", "podcast", "地端模型", "lora", "開源模型", "OCR", "indra_ocr", "Kromoss", "ComfyUI", "繁體中文"] language: "en" license: "CC BY 4.0" license_url: "https://creativecommons.org/licenses/by/4.0/" attribution: "when reusing or quoting, credit the author and link back to the original" --- # 🧪 GPU-Free and 15x Faster: 1.25M Parameter indra-ocr is a Local Savior | Local AI Lab This issue features indra-ocr, an ultra-lightweight OCR model with only 1.25M parameters that processes Traditional Chinese OCR at high speeds on CPU servers without needing a GPU, and Kromoss, an image generation model that can quickly generate clear Chinese text in ComfyUI in just 8 steps. Are you a backend engineer looking to protect customer privacy without paying for expensive cloud APIs? We will show you how to build a high-throughput document processing pipeline locally using the most resource-efficient architecture. --- ## Featured This Issue ### 1. datalab-to/indra-ocr: Ultra-lightweight, CPU-first, blazing-fast OCR model supporting Traditional Chinese and LaTeX formula conversion - **What it is**: This is an ultra-lightweight, CPU-first OCR model with only 1.25M parameters, licensed under OpenRAIL. It is specifically optimized for Traditional Chinese and English, allowing you to quickly convert scanned documents, receipts, or academic PDFs into Markdown or structured JSON formats. - **Use Cases**: - **Backend engineers building fully local document processing pipelines**: When you need to automatically recognize and convert scanned invoices and receipts containing Traditional Chinese and English into structured JSON format to fully protect customer privacy, this model eliminates any compliance worries about sending data to the cloud. - **Full-stack developers building full-text search for historical PDFs**: If you need to build a search index for a massive archive of internal scanned documents but have a tight budget and want to avoid expensive GPU resources, this model can be deployed directly on existing CPU servers for high-throughput text extraction. - **Researchers converting academic papers**: If you have scanned papers containing complex mathematical formulas that need to be converted into editable Markdown format, you can use the `--math` parameter to let the model automatically convert formulas into precise LaTeX syntax. - **Can it run**: The official README labels it as CPU-first. The model size is extremely compact, with only about 1.25M parameters, and the ONNX pipeline is only around 18 MB. According to official data, on a 4-core vCPU x86 machine, the median processing time per page is just 212 milliseconds, or 357 milliseconds with math formula mode enabled, requiring absolutely no GPU involvement. Available formats include safetensors and ONNX. - **Popularity**: It currently has a trending score of 255 on Hugging Face, with 3.1k downloads and 285 likes. - **Comparison**: According to the official model card and third-party benchmarks, its processing speed can reach 44 pages per second on a 32-core AMD EPYC server, which is up to 15 times faster than Tesseract 5 and 10 times faster than PaddleOCR (PP-OCR). However, third-party evaluations also note that its error rate may increase when dealing with extremely old or blurry scanned books. - **How to get it**: According to the official README, developers can install it directly via `pip install indra-ocr`. Run `indra-ocr paper.pdf` in the terminal to output Markdown to stdout, or run `indra-ocr scan.png --format json` to quickly get a JSON format containing lines, bounding boxes, and layout blocks. - **Links**: [Hugging Face Page Link](https://huggingface.co/datalab-to/indra-ocr) --- ### 2. lodestones/Kromoss: Bilingual ComfyUI image generation model for fast, 8-step local generation of clear Chinese text - **What it is**: This is a community full-finetuned version based on MOSS-Image-Base with around 5B parameters, licensed under Apache-2.0. Its standout feature is support for bilingual Chinese and English prompts, and it comes with an 8-step DMD/DMD2 distilled LoRA, significantly reducing the time required to generate high-quality images with Chinese text locally. - **Use Cases**: - **Frontend engineers quickly generating UI mockups and prototypes**: When you are designing web prototypes and need to quickly generate UI mockups with specific Chinese/English typography and structured layouts (such as flowcharts or kanban boards on a whiteboard), you can use JSON format to precisely control the placement of objects and text in the image. - **Indie hackers creating localized product promotional illustrations**: If you need to create promotional images with Traditional Chinese handwritten fonts for your local projects, such as a detailed illustration of a fridge magnet that says "Remember to buy milk and eggs!", this model can seamlessly integrate the Chinese text directly into the image, eliminating the need to manually add text using Photoshop. - **Can it run**: The official README lists the model size at around 5B parameters in safetensors format. According to a third-party GPU evaluation site, running single-batch inference at FP16 precision requires an estimated peak VRAM of about 11 GB, which can be easily handled by consumer-grade graphics cards like the RTX 4090 24GB. Additionally, you can quantize it to INT8 (requiring about 5.5 GB VRAM) or INT4 (requiring about 2.8 GB VRAM) to further lower the local VRAM threshold. - **Popularity**: It currently has a trending score of 162 on Hugging Face, with 5.4k downloads and 163 likes. - **Comparison**: The official model card notes that this model is a full-finetuned version of MOSS-Image-Base. Compared to the official MOSS-Image-SFT, Kromoss is a seamless drop-in replacement. Furthermore, it comes with an 8-step DMD/DMD2 distilled LoRA, shortening the generation process from the original 50 steps to just 8 steps, greatly improving local inference efficiency. - **How to get it**: According to the official README, when using it in ComfyUI, you need to place `kromoss_bf16.safetensors` into the `models/diffusion_models/` folder, and pair it with the `qwen3.5_9b_bf16.safetensors` text encoder and `flux2-vae_small.safetensors` VAE. To enable fast 8-step generation, load `8-steps-dmd-lora.safetensors` and set the KSampler to euler/simple, 8 steps, and CFG 1. - **Links**: [Hugging Face Page Link](https://huggingface.co/lodestones/Kromoss) --- ## Conclusion As seen from this issue's selection, local AI models are rapidly evolving in two directions: extreme lightweighting and high localization. Whether it is indra-ocr, which runs at high speeds on CPU servers without even needing a GPU, or Kromoss, which perfectly generates Traditional Chinese handwritten text, both demonstrate the open-source community's incredible creativity in building practical tools. If you happen to be facing challenges with document recognition or UI mockup generation, why not download these models to your machine this weekend and give them a try? See you in the next issue! --- ## About this article and its author Originally published on [Mark Ku's Tech Notes](https://blog.markkulab.net/en/local-ai-lab/local-ai-lab-2026-10-21) License: [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/) — when reusing or quoting, credit the author and link back to the original ### About the author **[Mark Ku](https://blog.markkulab.net/en/author/mark-ku)** — Software engineer - 10+ years as a software engineer - Built North-American e-commerce and AI SaaS subscription billing ### Free tools built by the author All of these are free to use: - [Free PDF Sign Tool](https://blog.markkulab.net/en/tools/pdf-sign): Online PDF sign tool — draw, type, or upload a signature, then drag, resize, and download. Everything runs in your browser; nothing is uploaded. - [VS Code Refactory](https://blog.markkulab.net/en/tools/refactory): Refactory is a VS Code refactoring extension: 34 actions plus a 37-rule code-smell inspection layer with a Code Health dashboard, across 18 languages, backed by 534 tests. It learns your repo's conventions: where interfaces live, where DI is registered, whether 'use client' belongs. It ranks files by git churn × complexity so you know what to fix first, and hands any smell to the Claude Code already on your machine. Free to use, and your source never leaves your computer. - [DB-Kit Database Manager](https://blog.markkulab.net/en/tools/db-kit): DB-Kit is a lightweight, cross-platform database manager built with Tauri + Rust + React. Manage MySQL, MariaDB, PostgreSQL, SQL Server, Oracle, SQLite, MongoDB, Redis, Kafka, Elasticsearch and RabbitMQ from one consistent interface: passwords encrypted in the OS keychain, SSH tunnels, full CRUD, a visual query builder, stacked multi-statement result sets, cross-connection data transfer, schema & data compare with sync SQL and schema snapshots, Excel / CSV import & export, visualized execution plans, ER diagrams, scheduled backups, SQL stress testing with p50–p99 latency percentiles, a 15-rule SQL review engine, Review & Run (AI review plus per-statement backups and an auto-generated rollback script), Kafka message browsing with monitoring & alerts, an SSH terminal with SFTP / FTP, Docker / Kubernetes, remote desktop (RDP / VNC / RustDesk), file / folder compare, stored procedure integration tests, six UI languages, a built-in AI assistant (a local CLI or any Anthropic / OpenAI-compatible API; natural-language SQL, AI review and tuning advice) and the dbk CLI. Free and open source (MIT), with installers for Windows, macOS and Linux. - [VS Code Super Mermaid](https://blog.markkulab.net/en/tools/super-mermaid): Super Mermaid is a VS Code extension for beautiful Mermaid diagrams out of the box: auto-colored live preview, mouse pan & zoom, high-res PNG / SVG export, 21 templates and multiple themes. Free and open source (MIT). - [React Super Mermaid](https://blog.markkulab.net/en/tools/react-super-mermaid): react-super-mermaid is an open-source React component library: render beautiful Mermaid diagrams with a single , with built-in colorful / sketch themes, pan & zoom, in-diagram search, and high-res SVG / PNG export. Lightweight, SSR-safe, fully typed. Free and open source (MIT). - [Jira / Confluence Super Mermaid](https://blog.markkulab.net/en/tools/jira-super-mermaid): An Atlassian Forge app: write Mermaid syntax directly inside a Jira issue or a Confluence page and get flowcharts, sequence diagrams, state machines and Gantt charts. 11 diagram types, SVG / PNG export, light and dark themes, full CJK support. Runs on Atlassian: your diagrams live in your own site and the app calls no third-party service. Free, coming soon to the Atlassian Marketplace. - [Mermaid Live Preview](https://blog.markkulab.net/en/tools/mermaid-preview): Write Mermaid in your browser, see it render instantly, and share the whole diagram as a single link. No sign-up, nothing uploaded to a server, and mermaid.live share links work as-is. - [React Intl Phone Number](https://blog.markkulab.net/en/tools/react-intl-phone-number): react-intl-phone-number is an open-source React component: framework-agnostic and antd-free, with E.164 in/out, a searchable flag / country-code dropdown, configurable validation levels (strict / mobile-strict / loose), themeable CSS, and i18n — phone logic powered by google-libphonenumber. Lightweight and fully typed. Free and open source (MIT). - [Uptime Kuma Cluster](https://blog.markkulab.net/en/tools/uptime-kuma-cluster): Turn single-node Uptime Kuma into a highly available cluster: OpenResty + Lua smart load balancing, shared MariaDB state, health checks and automatic failover, plus cluster-management REST APIs. One Docker Compose command to start. Free and open source (MIT). - [AI Podcast Cut](https://blog.markkulab.net/en/tools/ai-podcast-cut): Drop in a recording and it removes fillers and stutters, levels loudness segment by segment, and sends a second agent to review every cut. Cut points snap to word boundaries and zero crossings, every splice gets a fade, and sentence-end breaths are preserved. Desktop app for Windows, macOS and Linux. MIT licensed; the Windows installer bundles ffmpeg. - [AI Video Cut](https://blog.markkulab.net/en/tools/ai-video-cut): Pick something in your video: type it (face, license plate, phone screen, logo), click points or drag a box, or let Claude Code / Codex pick it. AI Video Cut tracks it frame by frame, then mosaics or blurs it, recolors it, pins stickers and text to it, swaps a screen or poster for your own image or video, or removes it using background that other frames actually captured. Plus sequence editing, local captions and an AI assistant. Free, open-source (MIT) Windows desktop app (macOS / Linux experimental) that runs locally, nothing uploaded. - [open-pos restaurant POS](https://blog.markkulab.net/en/tools/open-pos): One computer and one receipt printer is enough to open the shop. Your data lives on your own disk, no subscription, no lock-in, MIT licensed. Money is integer New Taiwan dollars with tax split by the statutory formula, so sales + tax always equals the total. Printing goes straight over ESC/POS on TCP 9100, with no vendor driver. Tauri + Rust + SQLite desktop app, v1.0 in development. - [.NET e-commerce API template](https://blog.markkulab.net/en/tools/dotnet-ecommerce-api): An e-commerce backend API template built on .NET 10: Controller → Service → Repository layering, Autofac convention-based registration, SqlSugar for MySQL / MariaDB and SQL Server, Redis cache and message queue, Hangfire jobs, JWT auth, and an extensible payment provider module (ECPay, Newebpay). Apache 2.0. - [Open Token Monitor team AI usage](https://blog.markkulab.net/en/tools/open-token-monitor): Collect token usage and equivalent cost from every employee's AI coding tools into a hub you host yourself: PostgreSQL, org roster, dashboards compared by company and department, and a reports API. Electron and Rust/Tauri clients upload every 30 minutes. MIT licensed, one Docker command. - [ai-security-monitor defensive security](https://blog.markkulab.net/en/tools/ai-security-monitor): ai-security-monitor is a single Rust binary for your computers and servers. It finds unusual connections, rates them safe / unknown / suspicious / dangerous against 26 public threat-intel sources, and checks firewalls, open ports, vulnerabilities and 60 hardening rules. Security Hub shows the whole fleet, and Claude Code / Codex explain it in plain words. MIT, free, self-hosted. ### Daily podcasts - [Mark's Tech Insights — Daily AI News](https://blog.markkulab.net/en/category/tech-news): Daily curated AI and tech trends. Catch the latest developments via audio summaries — covering AI applications, software architecture, DevOps, and engineering practice. — RSS: https://blog.markkulab.net/feed.xml - [AI股市蝦聊](https://blog.markkulab.net/en/category/ai-stock-chat): Every trading day, an AI-analyzed take on the Taiwan stock market, delivered as a two-host conversation covering the session and the next-day outlook. — RSS: https://blog.markkulab.net/ai-stock-chat/feed.xml - [開源好物週報](https://blog.markkulab.net/en/category/open-source-weekly): A weekly two-host pick of free open-source tools surfaced from real Hacker News, GitHub, and Reddit buzz — what pain they solve and the fastest way to get started. — RSS: https://blog.markkulab.net/open-source-weekly/feed.xml - [AI 運動週報](https://blog.markkulab.net/en/category/sports-weekly): Two hosts talk NBA, MLB and world sport three times a week — scores, records and the stories behind them, from a Taiwanese fan perspective. — RSS: https://blog.markkulab.net/sports-weekly/feed.xml - [AI 國際新聞快報](https://blog.markkulab.net/en/category/world-news): A daily two-host briefing that makes sense of the past 24 hours in world news — geopolitics, the global economy, conflict and security, disasters and climate — from a Taiwanese perspective, neutral and fully sourced. — RSS: https://blog.markkulab.net/world-news/feed.xml - [地端 AI 實驗室](https://blog.markkulab.net/en/category/local-ai-lab): Twice a week, a two-host look at open-weight models and LoRAs you can actually run on your own machine: what they are for, whether your GPU can handle them, and how they compare — sourced from official model cards. — RSS: https://blog.markkulab.net/local-ai-lab/feed.xml - [AI 行銷週報](https://blog.markkulab.net/en/category/ai-marketing): Every Tuesday, two hosts break down the week in AI for marketing, advertising, SEO/GEO and content marketing: practical, never sponsored, every story sourced. — RSS: https://blog.markkulab.net/ai-marketing/feed.xml - [AI 資安週報](https://blog.markkulab.net/en/category/ai-security): Every Thursday, two hosts cover the week where AI meets cybersecurity: LLM and AI agent vulnerabilities, prompt injection, model supply-chain risks and AI-powered attacks, from a defense-first angle with every story sourced. — RSS: https://blog.markkulab.net/ai-security/feed.xml ### Newsletter [Subscribe to the newsletter](https://blog.markkulab.net/en/subscribe) — Be the first to know about new posts. No spam, unsubscribe anytime.