Snowflake unveils Cortex Code AI agent for developers

Snowflake has expanded its Snowflake Cortex AI suite with the launch of Cortex Code, a dedicated AI coding agent engineered to operate within an organisation's specific data context and development frameworks.

The new tool is designed to assist with a broad spectrum of technical tasks, ranging from data engineering and analytics to machine learning and the construction of autonomous agents.

According to Snowflake, the agent is "Snowflake-native," meaning it possesses a deep architectural understanding of the platform's data, compute, governance, and operational semantics. Christian Kleinerman, Snowflake's EVP of Product, emphasised the shift towards operational utility, noting that for AI to provide genuine value, it must evolve from an experimental tool into a core component of everyday systems.

To ensure versatility, Cortex Code is compatible with both Snowflake's native user interface and local developer environments.

The company has integrated the tool with existing security and governance controls, aiming to provide a secure workspace for AI application development and team collaboration without compromising organisational data standards.

Product rollout

Snowflake said Cortex Code will be available in two forms. Cortex Code in Snowsight will be generally available soon. Cortex Code CLI is now generally available.

The company said Cortex Code CLI works in a terminal and integrates with code editors including VS Code and Cursor. Snowflake framed this as a way for teams to use Snowflake-aware assistance within familiar workflows rather than shifting into a separate environment.

Snowflake also said Cortex Code can support the full development lifecycle from initial design and implementation through optimisation and operations. The company said it is "customizable and interoperable". It did not outline commercial terms or pricing details in the announcement.

Customer adoption

Snowflake named a group of customers that it said already use Cortex Code. These included Braze, Decile, dentsu, FYUL, LendingTree, Shelter Mutual Insurance, TextNow, United Rentals, and WHOOP.

The company's pitch focuses on a familiar problem for data teams. Many AI coding assistants prioritise code generation, but organisations often have to balance speed with governance, accuracy, and operational constraints. Snowflake argued that the gap sits in an assistant's ability to work with enterprise data context and policies.

One customer statement pointed to governance and workflow fit. "Our teams operate in an industry where the demand for high-quality, data-driven marketing solutions is accelerating rapidly. To keep pace, we need tools that let us scale efficiently while maintaining consistency and governance," said Joe Tobey, Head of Data Products Engineering, dentsu.

"Cortex Code CLI aligns naturally with how our teams work, enabling them to translate data and evolving requirements into AI-powered solutions on Snowflake faster, supporting our ability to meet growing market expectations without disrupting established workflows," said Tobey.

Workflow focus

Other customer comments emphasised time from experimentation to production. "Cortex Code is fundamentally changing how our teams build on Snowflake," said Miks Lūsītis, Senior Director of Data, FYUL.

"By bringing context-aware AI directly into our development workflows, Cortex Code has helped us move from experimentation to production faster without having to switch between tools or question if the agent understands our business context," said Lūsītis.

LendingTree's comments focused on iteration cycles in data and analytics planning. "As we look at how agentic AI can accelerate our data and analytics roadmap, speed and iteration are critical," said Srinivas Madabushi, Senior Vice President, Technology, LendingTree.

"Cortex Code gives our teams a simple, in-platform way to move quickly from exploring ideas to delivering AI-driven workflows directly on Snowflake. It has the power to help us shape how we roll out AI-powered capabilities for more personalized consumer experiences and smarter financial decisioning," said Madabushi.

New integrations

Alongside Cortex Code, Snowflake announced an integration with v0 by Vercel. Snowflake said this integration will be generally available soon. It described the integration as a way to create AI-powered data applications using natural language and then deploy them inside Snowflake via Snowpark Container Services.

Snowflake also announced an integration with the Brave Search API, which it said is in public preview. The company said this adds real-time web information into Snowflake Intelligence, Cortex Code, and Cortex Agents. Snowflake said this links internal enterprise data with public web context and supports use cases such as questions on current events, market trend research, and documentation retrieval.

Workspace updates

Snowflake also announced updates to its Workspaces tooling. The company said Shared Workspaces and Snowflake Notebooks are now generally available. It also said OpenID Connect-based authentication is now generally available.

Shelter Mutual Insurance described a regulated environment requirement. "What stands out about Cortex Code is how naturally it fits into the way our teams already work," said Vibhor Gupta, Vice President of Enterprise Data & AI, Shelter Mutual Insurance.

"It helps us reduce friction in everyday data and AI development while maintaining the controls and oversight we need in a regulated environment. With Cortex Code, our teams can build faster with the context they need to be successful," said Gupta.

United Rentals highlighted performance work on business intelligence systems. "Snowflake Intelligence is already helping our teams make faster, better decisions across the business, and Cortex Code is extending that intelligence into the AI experiences we build for our team," said Tony Leopold, Chief Technology and Strategy Officer, United Rentals.

"Cortex Code helps our engineers improve the performance of our business intelligence tools, meaningfully reducing the time it takes to improve quality and speed of Natural Language Query responses," said Leopold.

Snowflake said it will continue to expand AI-powered development across its product stack, including Snowflake Intelligence, Cortex Code, and Cortex Agents.