Awesome GitHub RepositoriesLLM-Based Text Processing

openai-translator is a cross-platform translation tool and language learning utility available as a browser extension and desktop application. It functions as a client for large language model APIs to translate, summarize, and polish text across different digital environments. The project differentiates itself by integrating optical character recognition to translate text extracted from images and screenshots. It also includes a language learning workflow that allows users to save new vocabulary to a digital book and use text-to-speech synthesis for pronunciation. The tool provides broad tex

Nextai-translator is an AI-powered text processor and cross-platform translation application. Available as a desktop app and browser extension, it uses large language model APIs to translate, summarize, and refine multilingual content in real time. The tool integrates with clipboard managers and text selection utilities to trigger automated translations immediately after content is copied or highlighted. It also functions as an OCR translation utility, extracting and translating text from screenshots and non-selectable image content. Additional capabilities include a vocabulary management sy

ChatPaper is a suite of AI agents and utilities designed for academic literature automation, manuscript editing, and research assistance. The system functions as a research assistant that summarizes, translates, and analyzes scholarly papers, while providing specialized tools for converting academic PDFs into structured markdown to preserve formulas for analysis. The project features a literature survey automator that crawls research repositories and synthesizes domain reports, alongside a research mind map generator that transforms linear document content into non-linear node-based maps. It

anx-reader is a cross-platform e-book reader and cloud-synced library manager. It renders various electronic book formats into a standardized HTML view with customizable themes and fonts for a consistent experience across different operating systems. The project integrates a large language model as a reading assistant to summarize text and answer questions about book content. It also functions as a digital annotation tool for creating color-coded highlights and detailed notes for external research export. The system includes capabilities for organizing digital library collections, synchroniz

This project is a long context inference engine and optimizer designed to process infinite text streams using large language models without memory growth or performance degradation. It serves as a system for maintaining constant memory usage during the generation of text from arbitrarily long input sequences. The implementation utilizes a rolling key-value cache manager and attention sink mechanisms to stabilize the attention process during continuous stream processing. By retaining initial tokens and employing a sliding window of key-value pairs, the system enables constant-time inference an

Bili.Copilot, Bilibili के लिए एक नेटिव Windows डेस्कटॉप क्लाइंट है जो AI-आधारित मीडिया ब्राउज़िंग और वीडियो समरी के लिए लार्ज लैंग्वेज मॉडल्स (LLMs) का उपयोग करता है। यह एक मीडिया ब्राउज़र और वीडियो समराइज़र के रूप में काम करता है, जो सबटाइटल्स और टेक्स्ट को प्रोसेस करके वीडियो और आर्टिकल्स का संक्षिप्त विवरण तैयार करता है। यह एप्लिकेशन यूज़र्स को AI मॉडल्स के साथ इंटरैक्ट करने, विशिष्ट जानकारी खोजने और वीडियो कंटेंट का मूल्यांकन करने की सुविधा देता है। ये क्षमताएं Windows App SDK और WinUI का उपयोग करके बनाए गए नेटिव Windows इंटरफ़ेस के माध्यम से मिलती हैं। यह मीडिया मैनेजमेंट और कंटेंट कंजम्पशन को कवर करता है, जिसमें ऑफ़लाइन देखने के लिए वीडियो डाउनलोडिंग और इंटीग्रेटेड प्लेबैक रेंडरिंग शामिल है। साथ ही, यह QR-आधारित ऑथेंटिकेशन और वेब-लॉगिन के जरिए यूज़र आइडेंटिटी को भी मैनेज करता है।

Sparrow एक LLM डॉक्यूमेंट एक्सट्रैक्शन प्लेटफॉर्म और विज़न-आधारित इन्फरेंस इंजन है जिसे छवियों और PDFs को वैलिडेटेड स्ट्रक्चर्ड डेटा में बदलने के लिए डिज़ाइन किया गया है। यह एक एजेंटिक वर्कफ़्लो ऑर्केस्ट्रेटर के रूप में कार्य करता है जो वर्गीकरण, निष्कर्षण और सत्यापन कार्यों को मल्टी-स्टेप पाइपलाइनों में जोड़ता है। यह सिस्टम एक बैकएंड-अज्ञेयवादी इन्फरेंस लेयर के माध्यम से खुद को अलग करता है जो लोकल GPUs, Apple Silicon और क्लाउड प्रदाताओं के बीच मॉडल का प्रबंधन करता है। यह एक्सट्रैक्ट किए गए टेक्स्ट को सटीक बाउंडिंग बॉक्स निर्देशांकों पर मैप करने के लिए कोऑर्डिनेट-आधारित विजुअल ग्राउंडिंग का उपयोग करता है और ध्यान केंद्रित करने और डेटा फॉर्मेट को सामान्य करने के लिए हिंट-आधारित मॉडल स्टीयरिंग का उपयोग करता है। यह प्लेटफॉर्म डॉक्यूमेंट इंटेलिजेंस वर्कफ़्लो को कवर करता है, जिसमें संरचनात्मक अखंडता बनाए रखने के लिए विशेष इमेज-आधारित टेबल प्रोसेसिंग और एक्सट्रैक्ट किए गए फील्ड्स की शुद्धता को सत्यापित करने के लिए स्कीमा-आधारित सत्यापन शामिल है। यह API प्रदर्शन, उपयोग एनालिटिक्स और सिस्टम हेल्थ की निगरानी के लिए एक डॉक्यूमेंट एनालिसिस डैशबोर्ड भी प्रदान करता है। आर्किटेक्चर में इंडेक्सिंग और ऑर्केस्ट्रेशन के लिए उपयोग की जाने वाली थर्ड-पार्टी लाइब्रेरीज़ को एकीकृत करने के लिए एक प्लगइन-आधारित एक्सटेंशन सिस्टम शामिल है।

Humanizer is a text processing system designed to remove machine-generated patterns from writing to make it sound more natural and conversational. It functions as an auditor and rewriter that identifies robotic signatures, formulaic tropes, and mechanical formatting in machine output. The project features a style-matching system that analyzes provided writing samples to replicate a user's specific sentence rhythms, vocabulary, and punctuation habits. This allows the tool to mirror a personal voice and apply a calibrated tone to the rewritten text. The system covers a broad range of linguisti

RedInk is an AI content automation tool designed to generate coordinated social media posts, including titles, body text, and matching visual assets, from a single user-provided topic. It functions as a stateless content pipeline that uses large language models to transform topics into structured marketing copy and image prompts. The system utilizes prompt-template orchestration to combine static instructions with dynamic inputs, guiding artificial intelligence toward specific output formats. Users can manage these AI behaviors and API preferences through a web-based settings interface that t

CapsWriter-Offline is a suite of desktop tools that operates without an internet connection, combining local media browsing, voice dictation, audio and video transcription, and 360-degree media viewing into a single application. The project's core identity centers on providing offline functionality for both media handling and speech-to-text workflows. What distinguishes it is the integration of voice dictation with a persistent local storage layer that saves every audio recording and daily transcript logs, along with a rule-based text normalization engine that converts spoken number phrases a

Novel-Plus is a content management system and online reading platform designed for hosting, managing, and distributing web novels across PC and mobile web interfaces. It provides a comprehensive environment for web novel publishing, featuring a multi-platform reader with bookshelves, customizable themes, and community commenting systems. The platform integrates an automated content crawler to gather and update literary data from external remote sources and employs a scalable text storage system using database sharding and flat files to handle large volumes of content. It also includes an inte

This project is a Sina Weibo web scraper and social media data pipeline designed to extract user profiles, posts, comments, and multimedia assets. It functions as a containerized data crawler that automates the collection and local storage of social media content and engagement metrics. The system includes a processing layer that uses large language models to analyze scraped text, generating summaries and sentiment analysis. It differentiates itself through a deployment-ready container model featuring an HTTP interface for managing extraction tasks and monitoring job progress. The crawler co

This repository is a collection of node-based pipeline configurations, examples, and templates for generating AI media. It provides a workflow library and a curated gallery of blueprints designed for creating images, videos, and 3D assets using diffusion models. The project specifically offers a set of pre-configured node graphs for implementing advanced image generation and refinement techniques, with a focus on Stable Diffusion workflows. These examples demonstrate how to interconnect processing nodes to define complex generative logic without writing code. The available templates cover a

Humanizer-zh is a tool designed to transform AI-generated content into natural-sounding writing by removing repetitive patterns and mechanical formatting. It functions as an AI content de-optimizer and text humanizer specifically focused on making Chinese text sound more human-authored. The project specializes in Chinese language refinement, replacing corporate jargon and artificial linguistic markers with native expressions and varied sentence rhythms. It employs a system to strip formulaic signatures and repetitive structures common in large language model outputs to increase perceived auth