Merlin by Encord.

Merlin by Encord

Merlin is the agentic intelligence layer that transforms Encord into the first AI data infrastructure platform with native agentic capabilities. By embedding directly into the tools teams already use—starting with MCP, Claude, and Codex—Merlin eliminates manual setup and accelerates every stage of the data lifecycle through natural conversation. Teams can now build, observe, and optimize their AI data infrastructure without leaving their workflow, turning complex data operations into simple, intent-driven interactions.

Product Highlights

• Conversational Project Setup: Go from a simple prompt or document to a complete labeling setup instantly—Merlin defines label schemas, customizes interfaces, and configures review stages automatically, eliminating the blank-slate problem.
• Real-Time Data Observability: Get on-demand visibility into your data's health with instant metrics and coverage gap reporting, removing dependency on manual analysis passes.
• Intelligent Optimization: Automatically surface data issues blocking model performance and navigate directly to fixes within Encord, closing the loop between detection and resolution.
• Native Workflow Integration: Launch and operate entirely within familiar agentic coding platforms via MCP, with Slack and additional integrations coming soon—no context switching required.

Use Cases

• Rapid AI Development Cycles: Accelerate iteration by reducing data preparation from days to minutes, enabling teams to keep pace with increasingly short development cycles and conversational interfaces.
• Data-Centric AI Improvement: Continuously monitor and enhance dataset quality through automated gap analysis and guided fixes, directly addressing the root cause of model performance limitations.
• Cross-Functional Collaboration: Enable engineers, data scientists, and domain experts to interact with complex data infrastructure through natural language, lowering barriers to effective AI development.

Target Audience

Merlin is designed for AI teams and machine learning engineers at organizations building production-grade AI systems who need to move fast without sacrificing data quality. It serves data-centric AI practitioners who recognize that model performance depends on getting the data right—and want agentic intelligence to handle the heavy lifting across their existing tool stack.