May 2025 Summaries

The final day of Spring Launch Week introduces new Retrieval-Augmented Generation (RAG) admin controls aimed at providing developers, startups, and SaaS teams enhanced oversight of their applications, alongside a special offer of 3,000 bonus pages for those signing up for the Starter plan. The launch includes the Partition Limit API, which allows precise control over data usage by setting caps on pages processed and hosted per partition, with features like immediate enforcement, continuous monitoring, and monthly reset. Additionally, partition patterns utilizing glob-style syntax enable organized routing of webhook events for multi-tenant environments, allowing for workload separation by environment, feature, or team. These new features are available for all Ragie users, offering tools to simplify microservice architectures and enhance multi-team integrations, with guidance and best practices accessible through documentation and scheduled calls.

Base Chat, a newly launched cloud service building on an existing open-source product, offers an intuitive way to interact with company knowledge bases by using advanced multimodal retrieval technology from Ragie, similar to that used in projects like JFK Files and Mearsheimer.ai. This platform integrates with various data sources such as Google Drive, Notion, and JIRA, and allows users to choose from AI models by Anthropic, Open AI, and Meta, providing personalized experiences with custom prompts and avatars. Administrators can manage user interactions and create public chatbots for sharing knowledge externally. Ragie customers can connect their data with Base Chat, enabling a conversational interface for data queries and the development of custom applications. The service is priced at $18 per user monthly, with different data plans and optional audio and video features available on a pay-as-you-go basis.

Ragie has announced an expansion of its connector ecosystem to improve integration with customer support and enterprise file management systems for Retrieval-Augmented Generation (RAG) applications. This includes new connectors for Zendesk and Intercom, allowing for seamless import of help center articles with preserved metadata, which enhances AI-driven support tools with accurate information. Additionally, new integrations with SharePoint, Backblaze, and enhanced Google Drive capabilities allow users to upload and manage content more efficiently within enterprise environments. Ragie is committed to continually developing new connectors to facilitate data integration from both internal and external systems, inviting users to request specific connector needs through their support channel.

Ragie has launched a suite of new open-source developer tools and enhancements aimed at improving the development of Retrieval Augmented Generation (RAG) workflows. These tools include the ragie-mcp-server, which allows comprehensive access to an organization's knowledge base, and ragie-cli, a versatile command-line interface for managing data imports from various sources. Promptie serves as a test environment for AI prompt engineering, integrating Ragie's retrieval capabilities with Anthropic's generative AI to refine prompts and optimize retrieval settings. Base Chat offers a reference application for multi-tenant RAG chatbots, enabling interaction with knowledge bases through an intuitive interface. Additionally, the langchain-ragie integration simplifies the development of context-aware AI systems by connecting Ragie with LangChain, facilitating the creation of precise, knowledge-grounded applications. All these tools are available under the MIT license, allowing developers to freely use, modify, and share them.

Ragie introduces native multimodal support for audio and video content, enabling users to upload files, ask questions, and receive answers, all while streaming from the exact relevant moments in the media. This advancement is powered by A/V Retrieval Augmented Generation (RAG), which offers integration options like Google Drive for seamless synchronization of multimedia content. Ragie's advanced infrastructure processes audio and video through transcription, visual description, chunking, indexing, and retrieval, reducing engineering overhead and accelerating product development. The platform supports a wide range of audio and video formats and multiple languages, providing efficient and scalable search and streaming capabilities. These features enable various applications across industries such as corporate training, media, healthcare, legal, and customer support, allowing developers to easily build immersive multimedia generative AI solutions without complex infrastructure.

Summarie is an open-source application designed to summarize documents, utilizing technologies like Next.js 15 and the Ragie TypeScript client to allow users to upload various file types and obtain detailed summaries. The setup involves creating a Next.js project, installing the Ragie SDK, and configuring API routes to handle file uploads securely without exposing API keys. The application includes features such as a simple file upload interface, a backend API for processing files with Ragie, and a mechanism to track document status using polling and webhooks. Once documents reach a "ready" state, summaries are automatically generated and displayed. The app emphasizes best practices including file validation, error handling, and user experience enhancements like progress indicators and feedback on uploads. While the example serves document summarization, the Ragie platform supports broader AI use cases, and the complete code is available in an open-source repository on GitHub.