Back to Servers

Topsort MCP Server

Official

by Topsort

Ask Claude to 'show me which sponsored products are underpacing their daily budget across all active campaigns' and get a real-time answer pulled directly from your live retail media data — no dashboard switching, no CSV exports, no waiting for the analytics team's weekly report. Topsort MCP Server connects Topsort's auction-based retail media infrastructure to any MCP-compatible AI client. Launched on March 27, 2026, it's the first MCP server built specifically for retail media operations. Your AI agent gets direct read access to campaign performance metrics, auction data, delivery pacing, and cross-dimensional comparisons through Topsort's existing infrastructure. The practical value shows up in daily operations. A marketplace ad ops manager who currently spends 45 minutes each morning pulling reports from Topsort's dashboard can instead ask Claude: 'Which campaigns have a click-through rate below 1% but are still spending their full budget?' Natural language replaces dashboard navigation, and the MCP protocol means the response comes from live data, not a cached summary. Topsort serves major retailers including DoorDash, Poshmark, and Coles across 40+ countries. The company has raised $43.2 million with a $150 million post-money valuation, backed by investors including W23 Global (Tesco, Ahold Delhaize's VC fund). This isn't a side project — it's backed by the infrastructure that processes billions in commerce media. The MCP Server sits above Topsort's existing stack, so retailers don't need to rip and replace anything. If you're already a Topsort customer, enabling MCP access is a configuration change, not a migration. Compatible with Claude, GPT, Cursor, and other MCP-supporting clients. The limitation is clear: this is currently available only to existing Topsort customers by request. There's no public GitHub repo, no npm package you can install independently. You need a Topsort account and infrastructure already running. For non-Topsort shops, this is a preview of where retail media tooling is headed rather than something you can try today.

api integrationmcp-serverretail-mediacommercead-optimizationtopsort

Installation

# See GitHub for installation instructions

Key Features

  • Query live campaign performance data using natural language through Claude, GPT, or Cursor
  • Detect pacing and delivery issues across active campaigns in real time — no manual dashboard checks
  • Investigate performance changes with conversational queries like 'why did CTR drop 15% on campaign X yesterday'
  • Compare performance across dimensions: products, categories, time periods, regions — all via natural language
  • Supports agent-operated workflows: AI agents can monitor, analyze, and flag issues autonomously
  • Sits above existing Topsort infrastructure — no migration required for current customers

Use Cases

  • Morning performance review: ask Claude 'summarize all campaigns that underperformed their targets yesterday' instead of building custom reports
  • Budget pacing alerts: have an AI agent continuously monitor delivery rates and flag campaigns burning budget too fast or too slow
  • Cross-campaign comparison: 'Compare sponsored product CTR across DoorDash and Poshmark storefronts for Q1' — one prompt, one answer
  • Ad ops troubleshooting: 'Why is campaign X getting zero impressions since Tuesday?' — the agent checks auction data, targeting rules, and bid competitiveness
  • Executive briefings: generate weekly retail media performance summaries in natural language for non-technical stakeholders

FAQ

Server Stats

GitHub Stars
Updated
3/28/2026

Related Resources

Weekly AI Digest