Vertex AI, MCP, and a modernized platform

Vertex AI and platform updates

obseed modernized its AI and web platform to make analytics more reliable and provide a consistent foundation for new data tools.

AI and reliability

• Vertex AI: Existing language-model calls were consolidated into a central Google Vertex AI integration.
• Model Context Protocol: An authenticated MCP server enables standardized tool calls for AI-assisted analysis.
• Token quotas: An atomic quota system consistently tracks input, output, and thinking tokens across concurrent requests.

Platform and activity data

• Modernized interface: The web platform migrated to Tailwind CSS v4.1 with a more consistent shared design language.
• Operational reliability: Deployment workflows were hardened and monitoring workloads separated from primary application services.
• Activity data: Rowing activity classification and equipment management were extended for more precise training records.