VS Code + Data Workers: MCP Agents in the World's Most Popular Editor

VS Code MCP data engineering is now possible thanks to MCP extensions for the world's most popular editor. With Data Workers' 15 AI agents, VS Code becomes a data-aware IDE — schema introspection, lineage traversal, quality checks, and grounded SQL — without switching to Cursor, Claude Code, or another specialized tool.

VS Code MCP data engineering workflows are now possible thanks to the growing ecosystem of MCP extensions for Visual Studio Code. VS Code remains the most popular code editor in the world, with over 15 million monthly active users, and data engineers make up a significant portion of that user base. With MCP extensions and Data Workers' 15 AI agents, VS Code transforms from a code editor into a data-aware development platform — without requiring you to switch to a specialized AI IDE like Cursor or a terminal-based agent like Claude Code.

The appeal is simple: you keep the editor you already know, with all your extensions, keybindings, and workspace configurations intact, and you add MCP-powered data agents on top. No new tool to learn, no new subscription to manage, no disruption to your existing workflow. The agents connect to your Snowflake, BigQuery, dbt, and other data tools through MCP, and you interact with them through VS Code's chat interface or through extension-provided UI elements.

How MCP Works in VS Code

VS Code supports MCP through two primary paths. First, GitHub Copilot's agent mode (available to Copilot subscribers) supports MCP tool calling natively. Second, community extensions like Continue.dev and Cline provide MCP client capabilities independently of Copilot. Both paths let you connect Data Workers agents to your VS Code environment.

The underlying mechanism is the same regardless of which path you choose: an MCP client in VS Code sends tool call requests to Data Workers' MCP servers, which execute operations against your data infrastructure and return results. The client presents those results in VS Code's interface — as chat responses, inline annotations, or panel views depending on the extension.

Setting Up Data Workers MCP in VS Code

The setup process depends on which MCP client extension you use. For GitHub Copilot, configure MCP servers in VS Code's settings.json and enable agent mode in Copilot Chat. For Continue.dev, add the MCP server configuration to Continue's config file. For Cline, add the server in Cline's MCP settings panel. All three approaches point to the same Data Workers MCP server endpoint.

Regardless of client, the Data Workers side is identical: install with npm install -g @anthropic/data-workers, provide your warehouse credentials, and optionally configure your dbt project path. The agents start on demand when VS Code's MCP client invokes them. Our documentation covers the configuration for each VS Code MCP client with step-by-step instructions.

Data Engineering Workflows in VS Code with MCP Agents

Once connected, the Data Workers agents unlock several powerful workflows directly in VS Code:

Interactive schema exploration. Open the chat panel and ask about your data infrastructure. 'What tables contain customer data?' triggers the Catalog Agent to search your warehouse and return matching tables with column descriptions, row counts, and freshness indicators. The results appear in the chat with clickable links to explore further.

Grounded SQL writing. When you open a .sql file and start writing a query, the MCP client can invoke the Catalog Agent to provide schema context to the AI assistant. Instead of autocomplete suggestions based on generic patterns, you get suggestions grounded in your actual table and column names, types, and descriptions.

dbt model generation. Describe the model you need in the chat and the agents coordinate to generate it: Catalog Agent for schema context, Lineage Agent for upstream refs, Semantic Agent for business definitions, and Transformation Agent for the SQL. The generated model appears as a new file or inline edit that you can review and accept.

Pipeline debugging. Paste an error message from a failed dbt run or orchestrator alert. The agents trace through your pipeline, checking source freshness, schema changes, and test results to identify the root cause. The diagnosis appears in the chat with a suggested fix that you can apply directly.

Documentation generation. Point the agents at models with missing descriptions and the Documentation Agent generates column-level documentation based on data profiling, lineage analysis, and usage patterns. The descriptions are added to your schema.yml files as inline edits.

VS Code MCP Client Comparison

For data engineers who already pay for Copilot, enabling MCP agent mode is the fastest path. For teams that prefer open source tools or want to use local LLMs, Continue.dev offers the most mature open source MCP client experience. Cline is excellent for teams that want maximum control over agent interactions with a diff-based review flow.

VS Code Extensions That Complement MCP Agents

VS Code's extension ecosystem adds significant value on top of MCP agents. Several extensions pair particularly well with Data Workers:

• •dbt Power User. Provides dbt-specific features like model navigation, lineage visualization, and test running. Combined with Data Workers' Lineage Agent, you get both visual and programmatic access to your DAG.
• •SQLTools. Adds SQL execution capabilities directly in VS Code. Use Data Workers to generate queries and SQLTools to run them — all without leaving the editor.
• •Rainbow CSV. Makes CSV files readable with column highlighting. Useful when Data Workers agents return data samples for inspection.
• •YAML. Enhanced YAML support for schema.yml and dbt configuration files. Data Workers generates YAML that is immediately validated by this extension.
• •GitLens. Advanced git features including blame, history, and comparison. Useful for understanding who last modified a model when debugging issues surfaced by Data Workers agents.

VS Code vs Cursor for Data Engineering with Data Workers

The natural question is: why use VS Code with MCP extensions when Cursor offers a more integrated AI experience? The answer depends on your priorities. VS Code with extensions gives you more control over each component — you choose your MCP client, your LLM provider, and your AI interaction model. Cursor gives you a more polished, integrated experience where the AI is baked into every aspect of the editor.

For data teams, VS Code wins when: the team has invested heavily in VS Code configurations and extensions, the organization mandates specific tools and cannot adopt Cursor, the team wants open source MCP clients like Continue.dev or Cline, or individual engineers prefer to assemble their own tool stack. Cursor wins when: the team prioritizes AI-native editing and wants the tightest possible integration between AI and editor.

Getting Started with VS Code and Data Workers

Setting up Data Workers MCP agents in VS Code takes under ten minutes regardless of which MCP client you choose. Install Data Workers, configure your warehouse credentials, set up your preferred MCP client extension, and start asking questions about your data. The documentation includes setup guides for each VS Code MCP client. Visit the Product page to explore all 15 agents, or book a demo to see Data Workers running in VS Code against a live data stack.