AI Agents & Automation

Cell2Sentence (C2S) is an open-source framework designed for applying Large Language Models (LLMs) to single-cell transcriptomics. It implements the C2S-Scale framework, which transforms expression vectors into "cell sentences"—space-separated gene names ordered by descending expression. This innovative approach allows LLMs to natively model scRNA-seq data using natural language, unifying transcriptomic and textual data. The tool enables advanced single-cell tasks such as perturbation prediction, dataset summarization, cluster captioning, and biological question answering. C2S-Scale models, including those based on Pythia and Gemma-2 architectures, are available on Huggingface, with support for finetuning on custom prompt templates and multi-cell prompt formatting.

clinicalBERT is an open-source repository offering publicly available Clinical BERT embeddings, designed to advance clinical Natural Language Processing (NLP) research. It enables users to leverage pre-trained models like Bio+Clinical BERT and Bio+Discharge Summary BERT, which are finetuned from BioBERT or the cased version of BERT. The tool provides clear instructions for direct integration via the Hugging Face transformers library, simplifying access for researchers and developers. Additionally, it outlines steps to reproduce the pretraining process using MIMIC data and offers examples for downstream tasks such as Med NLI and NER, making it a comprehensive resource for those working with clinical text data.

CLIP_prefix_caption is an open-source image captioning model that provides a novel approach to generating descriptive captions for images. Unlike traditional methods that often require additional supervision like object annotation, this model only needs images and their corresponding captions for training, making it highly adaptable to various datasets. It leverages the powerful CLIP model for generating semantic encodings and fine-tunes a pretrained language model to produce meaningful sentences. The tool boasts significantly faster training times while maintaining state-of-the-art results, even on large datasets like Conceptual Captions. It also offers a variant using a transformer architecture for the mapping network, avoiding GPT-2 fine-tuning, and still achieving comparable performance on the nocaps dataset. The project provides inference notebooks and a GUI for easy visualization and use.

Claude Code Skill Factory is a powerful open-source toolkit designed for building and deploying production-ready Claude Skills, Code Agents, custom Slash Commands, and LLM Prompts at scale. It offers an interactive builder and pre-built commands to streamline the development process. Key capabilities include generating complete Claude Skills with proper formatting and documentation, creating specialized Claude Code Agents with enhanced YAML frontmatter, and generating mega-prompts for various roles with 69 professional presets. The toolkit also supports building Claude Code hooks for workflow automation with safety validation and language-specific templates, and creating custom slash commands with comprehensive validation. It further enables interoperability between Claude Code and OpenAI Codex CLI, making it a versatile solution for AI agent development.

Chatito is a powerful tool designed for generating datasets crucial for training and validating AI chatbot models and various Natural Language Processing (NLP) tasks. It utilizes a simple Domain Specific Language (DSL) to define and generate diverse training examples, supporting applications like named entity recognition and text classification. The project includes an online IDE, a DSL specification, an AST parser, and a generator implemented in TypeScript. Chatito helps prevent overfitting by providing a flexible way to generate data, acting as an intersection between data augmentation and a description of possible sentence combinations. It supports various output formats for popular NLU providers like Rasa, Flair, LUIS, and Snips NLU, and also offers a default format for custom models, allowing for custom entity arguments.

Curator is a scalable, open-source data preprocessing and curation toolkit developed by NVIDIA NeMo, designed to enhance the training of large language models (LLMs) and other AI models. It provides GPU-accelerated, modular pipelines for various data modalities including text, images, video, and audio. The tool supports a wide range of capabilities such as data deduplication, quality filtering, language detection, aesthetic filtering, NSFW detection, and ASR transcription. Curator is built to scale from individual laptops to multi-node clusters, leveraging NVIDIA RAPIDS™ libraries like cuDF, cuML, and cuGraph, along with Ray, to achieve significant performance improvements and cost reductions compared to CPU-based alternatives. It is a core component of the NVIDIA NeMo software suite, facilitating the entire AI agent lifecycle.

Databend is an open-source enterprise data warehouse built in Rust, offering a unified architecture for analytics, search, AI, and Python sandbox environments. It provides core capabilities such as large-scale analytics, vector search, full-text search, and auto schema evolution. Databend is agent-ready, featuring sandbox UDFs for agent logic, SQL for orchestration, transactions for reliability, and branching for safe experimentation on production data. Its architecture supports flexible agent orchestration with a control plane for resource scheduling, an execution plane for SQL orchestration, and a compute plane for isolated sandbox workers. Databend is cloud-native, elastic, and compatible with S3, Azure, and GCS, making it suitable for enterprise-scale AI workloads.

Rispose, formerly Unfetch, is an AI Agents & Automation tool designed to help businesses build and embed custom AI agents directly onto their websites or platforms. It enables automation of support, sales, and customer engagement through AI-powered assistants. Users can train their agents with up to 1,000 files, including PDFs, documents, and text files, and customize their behavior with specific instructions to match brand voice. The platform integrates with popular services like Shopify, WordPress, Notion, Wix, and Webflow. Rispose offers detailed history and metrics to track agent performance, understand user interactions, and facilitate continuous improvement. It provides a seamless and budget-friendly solution for integrating LLMs into existing web applications.

Trillo AI is an innovative platform designed to transform business requirements into full software blueprints, including detailed specifications, designs, and production-ready application code. Unlike traditional AI code generators, Trillo AI employs a multi-agent architecture with 16+ specialized agents that mirror expert team workflows, breaking down the specification process into reviewable steps. This human-in-the-loop approach allows users to comment on and regenerate any step's output, maintaining full control and ensuring accuracy. It excels in generating complex enterprise applications, providing 50-80% of the application code directly from the blueprint, which can be used as a working prototype or a solid starting point for development. The platform significantly reduces the time for specification and design from weeks to hours, fostering collaboration among architects, analysts, designers, project managers, and developers.

Amplify develops advanced AI Assistants specifically designed for the media world, aiming to revolutionize how media companies operate in the digital age. Their product suite includes Seiri, an all-in-one AI Assistant for extracting metadata, cataloging media, recognizing faces, categorizing content, generating summaries, and detecting objects. GeNews is their generative AI tool for storytelling, helping journalists create compelling video stories faster by automatically selecting visuals and assembling timelines. SeiriVoice offers real-time transcription, translation, and dubbing for multimedia content, enhancing efficiency and accuracy in multiple languages. Amplify focuses on optimizing traditional workflows from ingest to distribution, providing cutting-edge technologies and expertise to empower businesses in the media and entertainment industry.

Grably is a multi-modal human interaction data research company specializing in providing high-quality conversational and interaction datasets for AI development. They offer a wide range of data applications, including large-scale multilingual and multimodal datasets for LLM pretraining, low-resource language modeling, and multimodal model training. Grably also provides specialized datasets for embodied AI, robotics, long-form video analysis, audio/speech understanding, code intelligence, and scientific/technical domain modeling. Their process involves defining critical human activities, capturing synchronized multi-signal data, structuring it with precise annotation, and scaling to diverse populations. They also offer custom dataset design and delivery tailored to specific research, legal, and infrastructure requirements.

AI.Associates GmbH specializes in delivering comprehensive Artificial Intelligence and Machine Learning solutions and services designed to modernize data infrastructure and enhance business competitiveness. With over 25 years of expertise, their team offers AI consulting, solution development, and engineering services, focusing on Generative AI, AI Model Building, Feature Engineering, Hyper Parameter Tuning, AutoML, and MLOps. They help businesses transform data into actionable insights and build reliable, scalable AI solutions that integrate seamlessly with existing data infrastructure. Additionally, AI.Associates provides a range of AI & Machine Learning training courses, from fundamentals to customized programs, ensuring clients can effectively leverage AI technologies.

GODEL offers large-scale pretrained models specifically designed for goal-directed dialog, built on a Transformer-based encoder-decoder architecture. These models are trained for response generation grounded in external text, making them highly effective for dialog tasks requiring conditioning on external information, such as retrieved documents. The repository provides the dataset, source code, and pre-trained models, allowing for efficient fine-tuning on new dialog tasks with minimal task-specific data. GODEL V1.1, for instance, was trained on 551 million multi-turn dialogs from Reddit and 5 million instruction and knowledge-grounded dialogs, demonstrating improved performance, especially in zero-shot settings. It supports fine-tuning and evaluation across various dialog tasks and includes a demo interface for interaction.

GLM-TTS is a high-quality text-to-speech (TTS) synthesis system built on large language models, offering zero-shot voice cloning and streaming inference capabilities. Its two-stage architecture first employs an LLM to create speech token sequences, then uses a Flow model to convert these into high-quality audio waveforms. A key differentiator is its Multi-Reward Reinforcement Learning framework, which significantly enhances emotional expression and prosody, moving beyond the flat delivery of traditional TTS systems. It supports real-time audio generation, making it suitable for interactive applications, and offers multi-language support, primarily Chinese with English mixed text. The system also features phoneme-level modeling for fine-grained pronunciation control, addressing ambiguities in polyphones and rare characters through a 'Hybrid Phoneme + Text' input mechanism.

Hyperspec AI offers a generative real-time mapping solution, leveraging embodied AI with a cognition layer (LLM) for seamless voice and map integration. This technology is designed to enhance perception and catalyze autonomy at scale, particularly for applications requiring advanced computer vision, mapping, and sensor fusion. It aims to provide super-human perception capabilities, making it suitable for complex environments and dynamic scenarios where real-time spatial understanding is critical.

HCP-Diffusion is a comprehensive Diffusion model toolbox built on the RainbowNeko Engine, designed to simplify and unify Stable Diffusion workflows. It boasts a clean code structure and a flexible Python-based configuration system, making it ideal for conducting and managing complex experiments. The tool supports a wide array of training components and is highly extensible, flexible, and user-friendly compared to existing frameworks. Users can leverage a single Python config file to manage various training methods and model architectures, including Prompt-tuning (Textual Inversion), DreamArtist, Fine-tuning, DreamBooth, LoRA, and ControlNet. It also implements DreamArtist++, an upgraded version of DreamArtist based on LoRA, offering enhanced generalization, controllability, and faster training with minimal data.

aiXplain is an agentic OS designed for enterprise AI, offering a comprehensive platform to design, deploy, and govern mission-critical AI agents. It provides a full-stack solution with unified APIs, allowing for flexible development using code or no-code tools. The platform features an integrated marketplace with hundreds of LLMs, tools, and pre-built agents, supporting dynamic routing and RAG. aiXplain ensures no vendor lock-in, enabling seamless swapping of LLMs and tools. It supports deployment anywhere, including air-gapped and sovereign infrastructures, with auto-scaling, session isolation, and resilient execution. The platform also offers enterprise-grade governance with granular access controls, full audit visibility, centralized policy management, and built-in compliance enforcement.

Akridata provides an AI-powered visual inspection platform designed to optimize quality control in manufacturing and asset monitoring. The platform offers real-time defect detection using multimodal inputs like images and videos, seamlessly integrating into existing workflows. Key solutions include Vision Assist for AI-assisted human inspection, Vision Command for AI-driven oversight in manufacturing, and Vision Copilot for data science and model development. Akridata helps manufacturers reduce defect rates, ensure compliance, and accelerate model development, catering to industries such as automotive, medical, agriculture, and critical infrastructure. It enables continuous, accurate, and objective monitoring, eliminating human error and speeding up inspection processes.

gpt-fast is a highly efficient PyTorch-native transformer text generation tool, designed for minimal latency and a compact codebase of under 1000 lines of Python. It supports advanced features like int8/int4 quantization, speculative decoding, and tensor parallelism, making it suitable for high-performance applications. The tool is compatible with both Nvidia and AMD GPUs and is intended to showcase optimal performance achievable with native PyTorch, rather than serving as a comprehensive framework. Developers are encouraged to copy, paste, and fork the codebase for their specific needs, leveraging its efficiency for various LLM inference tasks.

APPLY is a Latvian company offering comprehensive AI solutions tailored for business growth. Their services encompass generative AI integration, where they identify areas for process improvement and develop customized systems. They also specialize in computer vision solutions for various industries, from medicine to forestry, providing consultations, business case studies, and project management. Furthermore, APPLY offers full-cycle robotics, mechatronics, and microscopy development, including high-performance optical and microscopic equipment. With over 50 AI projects completed and a team of 55 AI experts, APPLY positions itself as a leading AI partner in the Baltics, emphasizing a 100% execution rate and 14 years of experience.

Xinference, also known as Xorbits Inference, is a powerful and versatile library designed to serve language, speech recognition, and multimodal models. It simplifies the deployment and serving of both custom and state-of-the-art built-in models with a single command, making it accessible for researchers, developers, and data scientists. Key features include agent-native serving, automatic request batching for improved throughput, and distributed inference across workers. Xinference supports a wide range of models, including MiniMax-M2.7, GLM-5.1, Qwen3.6, and Gemma-4, and integrates seamlessly with popular third-party libraries like LangChain, LlamaIndex, Dify, and Chatbox. It offers flexible APIs, including OpenAI-compatible RESTful API, RPC, CLI, and WebUI, and intelligently utilizes heterogeneous hardware like GPUs and CPUs for accelerated inference.

KaibanJS is a JavaScript-native framework designed for building and managing multi-agent systems, leveraging a Kanban-inspired methodology. It allows users to create, visualize, and manage AI agents, tasks, tools, and teams, orchestrating AI workflows seamlessly. The framework provides real-time visualization of workflows, enabling users to track progress as tasks move through different stages, fostering more effective collaboration on AI projects. Key features include a Kaiban Board for visualizing agent workflows, role-based agent design for specialized AI agents, and robust tool integration supporting LangchainJS-compatible tools. It also offers sophisticated memory management, support for multiple LLMs, and a Redux-inspired state management architecture for consistent control across complex agent interactions. KaibanJS is designed for flexible integration into various JavaScript environments, including NextJS, React, Vue, Angular, and Node.js.

Keras-transformer is a Python library designed to facilitate the construction of (Universal) Transformer models within the Keras framework. It offers essential building blocks such as positional encoding, embeddings, attention masking, and memory-compressed attention. The library also supports Adaptive Computation Time (ACT) and provides a general implementation for BERT models, making it highly relevant for Natural Language Processing (NLP) tasks. Developers can flexibly piece together multi-step Transformer models using its Keras layers, or customize existing components like self-attention and activation functions. The repository includes practical examples demonstrating its application in language modeling with BERT and GPT on datasets like WikiText-2.

Kiln is a comprehensive platform designed to build, evaluate, and optimize AI systems, offering a suite of tools for developers and AI practitioners. It provides intuitive desktop applications for Windows, MacOS, and Linux, making advanced AI development accessible. Key functionalities include state-of-the-art evaluators for model quality, optimizers for prompts and models, and zero-code fine-tuning for various LLMs like Qwen, GPT, and Gemini. Kiln also supports Retrieval-Augmented Generation (RAG) for knowledge integration, agentic system building, synthetic data generation for datasets, and custom reasoning model training. Its open-source Python library and OpenAPI REST API facilitate integration into existing workflows, while its privacy-first design ensures local operation and data control.