AI Agents & Automation

Langcorn is an open-source API server designed to simplify the deployment of LangChain Large Language Model (LLM) applications and agents. Leveraging the high-performance FastAPI framework, Langcorn automates the serving process, making it easier for developers to operationalize their LLM solutions. Key features include easy deployment of LangChain models and pipelines, ready-to-use authentication functionality, and scalable architecture for language processing applications. It supports custom pipelines, asynchronous processing for faster response times, and provides well-documented RESTful API endpoints. Langcorn also allows for overriding default LLM parameters per request and handling memory for conversational AI applications, making it a versatile tool for LLMops.

Beam AI provides a leading platform for agentic automation, enabling enterprises to deploy self-learning AI agents for various operations. It helps reduce operational costs, increase efficiency, and transition organizations into AI-native market leaders. The platform supports building and deploying agents quickly, integrating with over 1000 systems including SAP and Salesforce. Beam AI agents are designed for enterprise scale and security, offering continuous learning, human-in-the-loop capabilities, and no-code development. It ensures secure, compliant, and auditable operations, meeting GDPR, ISO 27001, and SOC 2 Type II standards, with flexible deployment options including cloud and on-premise.

The open-trading-api GitHub repository offers comprehensive sample code for integrating with the Korea Investment & Securities (KIS) Open API. Designed for both Python developers and LLM-based automated trading environments, it simplifies the process of understanding and utilizing the KIS API. The repository includes functional unit samples for LLMs, practical API call examples for users, a strategy builder with visual UI, and a backtesting engine. It supports various API categories including domestic and overseas stocks, bonds, futures, options, ELW, and ETF/ETN. The tool aims to reduce development burden and facilitate the creation of AI-driven trading strategies and automated systems.

Fuel is an open-source data pipeline framework specifically designed for machine learning applications, developed primarily for use with Blocks, a Theano toolkit for training neural networks. It provides interfaces to common datasets like MNIST, CIFAR-10, and Google's One Billion Words, enabling users to easily access and manage diverse data sources. The framework supports flexible data iteration, allowing for minibatches with shuffled or sequential examples. A key feature is its pipeline of preprocessors, which facilitates on-the-fly data manipulation such as adding noise, extracting n-grams, or patching images. Fuel emphasizes serializability with pickle, ensuring that entire pipelines can be checkpointed and resumed for long-running experiments, relying heavily on the picklable_itertools library.

Funartech develops advanced AI solutions by integrating machine learning, operations research, and other fields to create novel algorithms for industrial problem-solving. The company emphasizes a human-in-the-loop approach, co-constructing solutions tailored to specific client needs, and boasts a 100% success rate. Their methodology often leads to significant industrial process improvements, ranging from 10% to 40%. Funartech also focuses on prescriptive analysis to enable optimal decision-making, and their solutions are designed for easy adaptability and integration into existing systems. They are actively involved in R&D, pushing the boundaries of AI to make robots emotive and intelligent, potentially paving the way for AGI.

TradingAgents-CN is a Chinese-enhanced version of the revolutionary multi-agent LLM-based financial trading framework, TradingAgents. Designed for learning and research, it helps users systematically understand how to use multi-agent trading frameworks and large AI models for compliant stock research and strategy experimentation. The platform does not provide real-time trading instructions, focusing instead on educational and research purposes. It features native OpenAI and Google AI support, custom endpoint configurations, intelligent model selection, and multi-LLM provider support. The tool also offers Docker containerized deployment, professional report export capabilities, and comprehensive support for A-share stocks, along with full Chinese localization.

JiwuChat is a lightweight, cross-platform instant messaging application built with Vue3 and Tauri2, offering a comprehensive suite of communication features. It integrates various AI assistants, including DeepSeek, Gemini, and Kimi, for enhanced chat experiences and AI group bots. Users can engage in real-time text, image, file, and voice messaging, as well as WebRTC-based audio and video calls with screen sharing. The app supports multi-device synchronization, ensuring seamless communication across Windows, MacOS, Linux, Android, and web platforms. JiwuChat also includes customizable themes, AI shopping functions, and translation tools, providing a versatile and efficient social networking experience.

unlock-deepseek is an open-source learning project dedicated to systematically interpreting and reproducing the DeepSeek series of AI models. It covers DeepSeek's advancements in large language models, mathematical reasoning, code generation, multimodal AI, inference models (like DeepSeek-R1), MoE architecture, and training infrastructure. The project aims to break down DeepSeek's cutting-edge technologies into understandable and reproducible learning content for a wide range of AI researchers and learners. Key features include in-depth paper analysis, hands-on tutorials for reproduction, technical breakdowns of core components, and comparative analysis with similar works.

HungaBunga is a powerful tool designed to automate the often time-consuming process of model selection in machine learning. It systematically brute-forces all scikit-learn models, including both classification and regression types, along with their respective hyperparameters. By running through every possible combination and ranking them using cross-validation, HungaBunga identifies the optimal model for a given dataset. This automation significantly streamlines the machine learning workflow, allowing data scientists and developers to quickly discover high-performing models without extensive manual experimentation. It supports various model categories like Generalized Linear Models, SVMs, Nearest Neighbors, Trees, and Ensemble methods, making it a comprehensive solution for model optimization.

Smart Web Assistant is a powerful Chrome extension designed to integrate advanced AI technology directly into your browsing experience. It provides real-time, context-aware insights on any web page you visit, leveraging either OpenAI's cloud API or local AI models like Llama or Mistral. This tool significantly boosts productivity and enhances understanding of web content across multiple languages. Key features include instant insights, multi-language support, customizable API settings, and secure storage of configurations within Chrome’s local storage. With up to five customizable shortcuts, users can quickly access frequently used queries, streamlining their workflow. It's an ideal tool for professionals, students, and anyone looking to make their internet experience more informative and productive.

FinGLM is an open-source project dedicated to building a robust and sustainable financial large language model. Its primary goal is to foster the integration of AI with finance through open collaboration and shared resources. The project offers a comprehensive framework for deep analysis of listed company annual reports, transforming complex financial texts into expert-level insights using AI. It addresses the significant challenges in real financial interactive scenarios by providing data preparation workflows, model fine-tuning processes, and a question-answering system. FinGLM also includes extensive datasets, such as 70GB of annual reports and 10,000 manually annotated evaluation data points, along with learning tutorials for data preprocessing, database usage, GLM, prompt writing, and model fine-tuning.

Elook.AI is an AI-powered platform designed for social mention monitoring and lead generation. It enables users to discover relevant discussions on Reddit, score potential leads based on engagement, and generate AI-powered response drafts to streamline communication. The tool aims to help businesses identify new opportunities, engage with their target audience effectively, and automate parts of their sales and marketing outreach. By focusing on social listening and AI-assisted lead nurturing, Elook.AI provides a comprehensive solution for enhancing customer engagement and driving growth.

OneClick Technologies provides AI-powered solutions designed to boost customer service efficiency and enhance customer experiences. Their suite of products, including OnePulse, Replic8, and OneInsight, offers advanced conversational and knowledge resolution tools. These tools help businesses reduce Average Handle Time (AHT), cut agent training and query escalation costs, optimize operations, and improve first-call resolution rates. OneClick Technologies focuses on delivering personalized and simplified customer interactions through AI-driven solutions, enabling seamless information transfer, automating query resolution, and increasing overall business productivity. Their offerings are tailored to transform customer engagement and operational efficiency for enterprises globally.

Aki is an AI-powered executive assistant designed to enhance personal and team productivity by streamlining task management and automating daily routines. It proactively supports users with tasks and calendars, allowing for instant planning by forwarding emails to Aki. The tool automates workflows, enabling users to delegate processes with simple prompts. Aki is built on top of Akiflow, a comprehensive productivity suite, and offers features like a meeting assistant, AI workflows, and a memory layer for a limitless second brain. It aims to improve life by acting as a proactive productivity coach and can even handle conversations on your behalf via phone, text, and emails.

BOND v1 is a Chrome extension designed to optimize meetings by automating key tasks. It provides real-time transcription of discussions, generates smart notes, and drafts follow-up emails, allowing users to concentrate fully on the conversation. This tool streamlines post-meeting workflows by capturing important information and facilitating effective communication. With customizable email templates, BOND v1 helps users manage team meetings, client interactions, and brainstorming sessions more efficiently, ensuring that key decisions and action items are documented and communicated promptly.

llm_related is an open-source GitHub repository dedicated to the implementation and study of various algorithms and concepts related to large language models (LLMs). The project serves as a learning record and a practical resource for developers and researchers interested in LLM development. It includes implementations for diverse topics such as DPO, GRPO, PPO, knowledge distillation, agent frameworks (langgraph_agent, llm_agent_zero), RAG demos, and training LLMs from scratch, including multimodal models and MoE architectures. The repository is primarily written in Python, making it accessible for those looking to understand and replicate advanced LLM techniques.

Paravision is a leader in trusted Identity AI, offering a suite of ethically developed AI building blocks designed for the next generation of identity verification. Its core products include ultra-accurate face recognition, capable of performing across diverse environments, demographics, and user angles. The platform also features advanced liveness detection to confirm user authenticity and detect spoofs, alongside deepfake detection to combat fraud and misinformation in identity and media. Additionally, Paravision provides age estimation for applications in age verification, retail, gaming, and digital media. The solutions are optimized for various deployments, including cloud, edge, mobile, or hybrid, and are trusted by world-leading service and solution providers across sectors like air travel, government programs, payments, and physical security.

EmailCrafted is an AI-powered email writing tool designed for professionals to generate personalized and professional emails in seconds. It allows users to customize the tone, structure, and brand voice of their communications. The platform offers flexible email solutions for diverse needs, including business communication, client outreach, and follow-up messages. EmailCrafted streamlines hiring processes, ensures consistent customer support responses, and helps build a strong brand presence through polished email communication. Users can register their brand voice by uploading an existing email or inputting key details, then simply enter recipient information and desired tone to generate a ready-to-send email.

ChillTrip is an AI-powered travel planner designed to simplify trip organization and sharing. It transforms your travel ideas into detailed, customizable itineraries, allowing you to plan, share, and explore trips with friends instantly. The platform focuses on a calm and structured planning experience, enabling users to input dates, budget, and desired vibe to generate a day-by-day plan. Users can then reorder days, pin activities, add notes, and share a trip link that opens directly in the app. ChillTrip also facilitates calm collaboration, keeping everyone aligned with clear timing, costs, and quick tips, and allows for previewing shared trips before importing them into the app.

Reclaim is an AI-powered calendar application designed to optimize individual and team schedules by intelligently managing focus time, meetings, and breaks. It integrates seamlessly with Google Calendar and Outlook, offering features like AI Focus Time to defend dedicated work blocks, AI Habits for recurring routines, and AI Smart Meetings to find optimal times across attendees. The tool also provides AI Scheduling Links for sharing availability, AI Buffer Time for breaks and travel, and AI Time Tracking for analyzing productivity. Reclaim is suitable for individuals, teams, and organizations, with pricing plans tiered by team size, and offers integrations with popular tools like Slack, Asana, ClickUp, Todoist, Jira, Linear, and Zoom.

CareerGenie is an AI-powered recruiting platform designed to significantly boost the capacity of fast-growing teams, talent agencies, and startups. It automates the entire candidate sourcing and screening process, from defining roles and creating job pages to discovering high-fit candidates across millions of profiles. The platform conducts AI-led interviews 24/7, providing detailed transcripts and clear signals to generate a ranked shortlist of top candidates. This allows human recruiters to focus on high-stakes conversations and final interviews, rather than tedious search, outreach, and initial screening tasks. CareerGenie aims to provide instant recruiting capacity, deeper candidate insights beyond resumes, and faster hiring decisions without increasing headcount.

Xandidate redefines the hiring process by leveraging AI to classify candidates precisely based on job descriptions. It trains a unique AI for each company job, ensuring bespoke matching. This platform transforms recruitment with intelligent candidate matching, automated screening, and data-driven insights. Key features include AI-powered candidate matching, automated screening, AI-optimized job descriptions, and bias reduction. It aims to reduce time-to-hire by 60% and improve hire quality, making it ideal for HR professionals and recruitment managers looking to streamline their hiring workflows.

Fin-R1 is a specialized large language model designed for complex financial reasoning, developed through a collaboration between SUFE-AIFLM-Lab at Shanghai University of Finance and Economics and FinStep.AI. Built upon the Qwen2.5-7B-Instruct base model, Fin-R1 undergoes a two-stage training process involving supervised fine-tuning (SFT) and reinforcement learning (RL) using high-quality, verifiable financial data. This approach enables it to achieve state-of-the-art performance across various financial benchmarks, including financial code generation, calculations, security compliance, smart risk control, and ESG analysis. The model is lightweight at 7B parameters, making it cost-effective to deploy while providing robust support for core financial business scenarios in banking, securities, insurance, and trusts.

The ChatGPT File Uploader Extension is a Chrome extension designed to significantly enhance the ChatGPT experience. It enables users to upload and process various text files, including .txt, .md, .js, .py, .html, .css, .json, and .csv, directly within the chat.openai.com interface. A key feature is its ability to automatically split large files into smaller, manageable chunks, preventing the ChatGPT model from being overloaded. This functionality is invaluable for tasks such as summarizing lengthy documents, analyzing data, finding bugs in code, and extracting insights from meeting notes or interviews. Users can adjust the 'Character Count' setting to prevent errors and ensure optimal processing. A progress bar keeps users informed about the upload status, making it a practical tool for anyone looking to extend ChatGPT's capabilities with external data.