お知らせ • Jun 03
Snowflake Announces Major New Capabilities For Snowflake CoCo And Introduces Snowflake Datastream
Snowflake announced major new capabilities for Snowflake CoCo (formerly known as Cortex Code), the coding agent where you build faster, making it easier for builders to automate workflows, develop apps, and operationalize AI on enterprise data through a simple prompt. Snowflake also introduced Snowflake Datastream, a new fully managed streaming service for Apache Kafka that brings real-time data and AI together in a single, governed platform, giving organizations an easier way to power AI apps and agents with fresh, continuously flowing data. CoCo makes it easier for builders to orchestrate complex data workflows, with an AI coding agent that automates the work for them. As a core part of Snowflake's agentic control plane for enterprise AI, CoCo gives builders a unified, governed environment to manage workflows across data, models, and apps. New capabilities in CoCo bring the experience across desktop, mobile, and Slack interfaces, meeting builders where they work to build, iterate, and ship without friction. In addition, builders can now let CoCo run tasks autonomously, assist with end-to-end app development and deployment, and help teams deliver production-ready results faster and more securely. Together, CoCo and Datastream give organizations a simpler way to build real-time AI apps by combining AI-assisted development with fresh, continuously flowing data in a single platform. What sets CoCo apart is that it's deeply integrated with Snowflake’s governed data platform, giving builders an AI agent that understands enterprise data, workflows, governance, and business context out-of-the-box. Available directly within the Snowflake platform and in the tools and environments where builders already work, CoCo makes building AI faster, easier, and more powerful. With new enhancements, even more builders can use CoCo to: Work with CoCo where teams already build: Builders can now harness CoCo where they work. Now available as a native desktop app, CoCo Desktop delivers a familiar, full-featured development experience deeply integrated with Snowflake data, governance, and AI workflows. In addition, builders can access CoCo through a new mobile app experience with CoCo Mobile App and through Slack with CoCo Slackbot, letting teams kick off workflows, check on tasks, and get insights on the go. CoCo is also extending into popular developer tools with a VS Code Extension, MicrosoftExcel Extension, and CoCo plugin for Claude Code to make building faster with the tools teams already use. Execute work autonomously and securely: CoCo can now execute tasks autonomously, eliminating the need for continuous user prompting. Automations power recurring, event-driven workflows for ongoing monitoring, validation, and operational processes. Every automation is secured through Snowflake’s role-based access controls and fully supported by comprehensive audit trails. New Cloud Agents enables users to start work in Snowsight and have it run securely in the cloud, without having to keep anything running locally on a laptop. It lets agents run in the background and return results when finished, so users don't need to stay in-session waiting. With the CoCo Agent SDK, builders can embed Cloud Agents capabilities directly into their existing workflows. A new secured local sandbox further ensures agents operate in an isolated local environment, protecting sensitive files and system resources so teams can automate confidently. Share what works across teams to build faster: CoCo now includes pre-built Skills for common data engineering and AI workflows across all data, while giving teams the flexibility to create and extend their own solutions. This includes CoCo Skills that help users build, operate, and troubleshoot data engineering workflows across the Snowflake ecosystem, from ingestion and transformation to orchestration and more. A new Skill Catalog makes it easy to discover, share, and reuse proven workflows across teams, helping turn individual solutions into resources everyone can use to boost productivity. Accelerate app delivery through ecosystem integrations: Developers can now turn simple conversations into full-blown apps in Vercel and deploy them directly into their Snowflake account on Snowflake App Runtime, where apps run inside the security perimeter with governance and access control inherited automatically. New integrations with Retool and Superblocks make it easier for organizations to build AI-powered apps on governed Snowflake data using the platforms their teams already know. Snowflake Datastream is built natively on Snowflake and fully Kafka-compatible, enabling organizations to connect existing streaming apps and data flows without rearchitecting their systems. Organizations get the openness and flexibility of standard Kafka for streaming ingestion together with Snowflake-native governance, security, and management in a single service. Teams can stream data directly into Snowflake tables without managing a separate streaming system alongside their data platform, while streaming data automatically inherits the same governance, access controls, masking, and lineage as the rest of the Snowflake platform. The result is a single, trusted foundation for real-time AI with significantly reduced operational complexity. Together, CoCo and Datastream simplify how organizations build and operationalize real-time AI. Datastream serves as the real-time engine powering fresh, continuously flowing data, while CoCo acts as the AI coding agent for building on top of it. Coupled with a new CoCo integration, builders can now use simple prompts to create and operationalize real-time pipelines and build AI-powered apps on top of live streaming data directly within Snowflake.