Together AI Integrations

Vertesia is an all-encompassing low-code platform for generative AI that enables enterprise teams to rapidly create, deploy, and oversee GenAI applications and agents at a large scale. Designed for both business users and IT specialists, it streamlines the development process, allowing for a smooth transition from the initial prototype stage to full production without the burden of extensive timelines or complex infrastructure. The platform supports a wide range of generative AI models from leading inference providers, offering users the flexibility they need while minimizing the risk of becoming tied to a single vendor. Moreover, Vertesia's innovative retrieval-augmented generation (RAG) pipeline enhances the accuracy and efficiency of generative AI solutions by automating the content preparation workflow, which includes sophisticated document processing and semantic chunking techniques. With strong enterprise-level security protocols, compliance with SOC2 standards, and compatibility with major cloud service providers such as AWS, GCP, and Azure, Vertesia ensures safe and scalable deployment options for organizations. By alleviating the challenges associated with AI application development, Vertesia plays a pivotal role in expediting the innovation journey for enterprises eager to leverage the advantages of generative AI technology. This focus on efficiency not only accelerates development but also empowers teams to focus on creativity and strategic initiatives.

LFM2 is a cutting-edge series of on-device foundation models specifically engineered to deliver an exceptionally fast generative-AI experience across a wide range of devices. It employs an innovative hybrid architecture that enables decoding and pre-filling speeds up to twice as fast as competing models, while also improving training efficiency by as much as threefold compared to earlier versions. Striking a perfect balance between quality, latency, and memory use, these models are ideally suited for embedded system applications, allowing for real-time, on-device AI capabilities in smartphones, laptops, vehicles, wearables, and many other platforms. This results in millisecond-level inference, enhanced device longevity, and complete data sovereignty for users. Available in three configurations with 0.35 billion, 0.7 billion, and 1.2 billion parameters, LFM2 demonstrates superior benchmark results compared to similarly sized models, excelling in knowledge recall, mathematical problem-solving, adherence to multilingual instructions, and conversational dialogue evaluations. With such impressive capabilities, LFM2 not only elevates the user experience but also establishes a new benchmark for on-device AI performance, paving the way for future advancements in the field.

The premier open-source AI assistant allows you to design personalized autocomplete functions and chat interfaces by seamlessly linking various models to diverse contexts. By eliminating obstacles that disrupt your productivity during software development, you can maintain a smooth workflow. With a user-friendly plug-and-play system that integrates effortlessly into your entire tech stack, you can expedite your development process. Customize your code assistant to adapt as it gains new features and enhancements. Effortlessly autocomplete entire blocks of code or individual lines across multiple programming languages as you continue typing. You can inquire about files, functions, the full codebase, and more by associating relevant code or context. Additionally, highlight specific segments of code and utilize a keyboard shortcut to transform that code into easily understandable natural language, enhancing collaboration and comprehension. This innovative tool empowers developers to work more efficiently and creatively than ever before.

gpt-oss-20b is a robust text-only reasoning model featuring 20 billion parameters, released under the Apache 2.0 license and shaped by OpenAI’s gpt-oss usage guidelines, aimed at simplifying the integration into customized AI workflows via the Responses API without reliance on proprietary systems. It has been meticulously designed to perform exceptionally in following instructions, offering capabilities like adjustable reasoning effort, detailed chain-of-thought outputs, and the option to leverage native tools such as web search and Python execution, which leads to well-structured and coherent responses. Developers must take responsibility for implementing their own deployment safeguards, including input filtering, output monitoring, and compliance with usage policies, to ensure alignment with protective measures typically associated with hosted solutions and to minimize the risk of malicious or unintended actions. Furthermore, its open-weight architecture is particularly advantageous for on-premises or edge deployments, highlighting the significance of control, customization, and transparency to cater to specific user requirements. This flexibility empowers organizations to adapt the model to their distinct needs while upholding a high standard of operational integrity and performance. As a result, gpt-oss-20b not only enhances user experience but also promotes responsible AI usage across various applications.

gpt-oss-120b is a reasoning model focused solely on text, boasting 120 billion parameters, and is released under the Apache 2.0 license while adhering to OpenAI’s usage policies; it has been developed with contributions from the open-source community and is compatible with the Responses API. This model excels at executing instructions and utilizes various tools, including web searches and Python code execution, which allows for a customizable level of reasoning effort and results in detailed chain-of-thought outputs that can seamlessly fit into different workflows. Although it is constructed to comply with OpenAI's safety policies, its open-weight nature poses a risk, as adept users might modify it to bypass these protections, thereby prompting developers and organizations to implement additional safety measures akin to those of managed models. Assessments reveal that gpt-oss-120b falls short of high performance in specialized fields such as biology, chemistry, or cybersecurity, even after attempts at adversarial fine-tuning. Moreover, its introduction does not represent a substantial advancement in biological capabilities, indicating a cautious stance regarding its use. Consequently, it is advisable for users to stay alert to the potential risks associated with its open-weight attributes, and to consider the implications of its deployment in sensitive environments. As awareness of these factors grows, the community's approach to managing such technologies will evolve and adapt.

Wan 2.6 is Alibaba’s flagship multimodal video generation model built for creating visually rich, audio-synchronized short videos. It allows users to generate videos from text, images, or video inputs with consistent motion and narrative structure. The model supports clip durations of up to 15 seconds, enabling more expressive storytelling. Wan 2.6 delivers natural movement, realistic physics, and cinematic camera behavior. Its native audio-visual synchronization aligns dialogue, sound effects, and background music in a single generation pass. Advanced lip-sync technology ensures accurate mouth movements for spoken content. The model supports resolutions from 480p to full 1080p for flexible output quality. Image-to-video generation preserves character identity while adding smooth, temporal motion. Users can generate complementary images and audio assets alongside video content. Multilingual prompt support enables global content creation. Wan 2.6 offers scalable model variants for different performance needs. It provides an efficient solution for producing polished short-form videos at scale.

The Nemotron-3 Super stands out as a groundbreaking addition to NVIDIA's Nemotron 3 series of open models, designed specifically to support advanced agentic AI systems capable of reasoning, planning, and executing complex multi-step workflows in challenging settings. It incorporates a distinctive hybrid Mamba-Transformer Mixture-of-Experts architecture that combines the streamlined capabilities of Mamba layers with the contextual richness offered by transformer attention mechanisms, enabling it to effectively handle long sequences and complicated reasoning tasks with notable precision and efficiency. By activating only a selected subset of its parameters for each token, this design greatly improves computational efficiency while ensuring strong reasoning skills, making it particularly suitable for scalable inference in demanding situations. With an impressive configuration of around 120 billion parameters, of which approximately 12 billion are engaged during inference, the Nemotron-3 Super significantly enhances its capacity for managing multi-step reasoning and facilitating collaborative interactions among agents in broad contexts. This combination of features not only empowers it to address a wide array of challenges in the AI landscape but also positions it as a key player in the evolution of intelligent systems. Overall, the model exemplifies the potential for future innovations in AI technology.

Scribe functions as an automated knowledge base on a self-hosted platform, leveraging your tools for content creation. It meticulously analyzes Git history, Claude Code and Codex sessions, as well as self-sent URLs and drop files, converting this data into a well-structured, cross-project wiki formatted in Markdown and preserved within Git repositories. By alleviating the necessity for developers to maintain an extra cognitive framework or to reconstruct context at the start of each agent session, Scribe adeptly captures essential decisions, solutions, assessments, and the reasoning behind them, allowing agents to access this accumulated knowledge before taking any actions. Its operational pipeline is orchestrated through cron jobs, which enables it to recognize projects and filter out low-value data using FTS5 before interacting with an LLM, thereby extracting validated facts through structured bounded workflows executed in two stages. The resulting output is organized into entity-centric pages that include YAML frontmatter, wikilinks, backlinks, retrieval context, and a variety of relationship types, such as supersedes, contradicts, derived_from, specializes, and extends, which ensures thorough documentation and clarity across various projects. This organized methodology not only fosters enhanced collaboration and knowledge sharing among teams but also streamlines the entire development process, ultimately leading to more efficient project management. As a result, teams can work more cohesively, drawing on a shared pool of insights and decisions that improve overall productivity.

Adaption offers users a robust suite of tools designed for the development of AI models that can adapt to changing data and evolving requirements. This all-encompassing web platform features vital functionalities like data preparation, synthetic data creation, dataset exploration, and model training processes. The Adaptive Data feature enables users to seamlessly import diverse files and datasets from platforms such as Hugging Face or Kaggle, while providing capabilities for column mapping, deduplication, prompt transformation, augmentation, translation, and evaluation. Through the Invent a Dataset function, users can create instruction or preference data tailored to specific written goals. The Visualize component further enriches the user experience by supporting inquiries with interactive charts and allowing for detailed examination of both text and tabular data at a granular level. AutoScientist is instrumental in refining training data and customizing model recipes to achieve targeted objectives, facilitating experimentation, outcome assessment, and the export of enhanced models. The platform also encompasses workflows that address fine-tuning, alignment, multilingual dataset expansion, and comprehensive data-coverage analysis, ensuring thoroughness in development. Moreover, Adaption provides API access alongside a Python SDK, making it easier to integrate these tools into existing AI development pipelines, thereby significantly boosting productivity and operational efficiency in the creation of models. In conclusion, Adaption stands out as an invaluable resource for developers aiming to effectively harness the power of AI technology in their projects.