Fodda Brand Intelligence

What Fodda Brand Intelligence MCP server does

Fodda Brand Intelligence MCP server gives MCP-compatible AI clients access to expert-curated knowledge graphs covering PSFK domains and partner material. Available subject areas include retail, beauty, sports, fashion, consumer electronics, and food and beverage. Additional sources include institutional reports, specialist industry graphs, company earnings calls, and macroeconomic context from up to 10 institutional data sources.

The server is intended for research and strategic analysis that benefits from traceable source material. It can return trend evidence, articles, citations, provenance, quantitative indicators, expert commentary, and relationships between concept nodes. The brand_tracker tool combines information across graphs into a Brand Intelligence Profile, while brainstorm_topic maps adjacent ideas and unexpected connections.

How it works

Clients connect to the MCP endpoint and discover the available schemas through the standard tools/list method. Search can be performed across selected graphs or across PSFK domain collections. Deterministic graph operations retrieve nodes, labels, neighbors, and evidence; search and synthesis tools handle semantic discovery and interpretation.

Long-running work is split into launch and status operations. deep_research_topic combines the knowledge graphs with live web search, and check_research_status returns progress or the completed narrative. The supplemental context tools use the same pattern. Synthetic analyst personas can be listed and then consulted for specialized synthesis. generate_visual produces branded SVG visualizations such as Cultural Shifts, Competitive Compass, and Innovation Pathway formats.

Setup and configuration

Hosted connections use a personal URL created from the Fodda Connections dashboard, or the HTTP endpoint at https://mcp.fodda.ai/mcp with OAuth or a Bearer API key. API keys may also be passed in the X-API-Key header. Legacy query-string authentication is deprecated. Client-specific setup guidance is provided for Claude Web, Claude Code, Claude Enterprise, ChatGPT, OpenAI Responses API, Cursor, Gemini, and custom agents.

For local deployment, install dependencies, build, and start the Node.js application:

• npm install
• npm run build
• npm start

The documented configuration variables are PORT, FODDA_API_URL, FODDA_MCP_SECRET, and NODE_ENV. The default HTTP port is 8080, the default upstream API is https://api.fodda.ai, and the default environment is production. Deployment instructions also cover Docker, Google Cloud Run, Kubernetes, and Terraform.

Tools and capabilities

The Fodda Brand Intelligence MCP server exposes tools for:

• Finding graphs, valid labels, trends, statistics, insights, and related nodes.
• Retrieving node metadata and source evidence with citations and provenance.
• Searching domain, expert, institutional report, and earnings intelligence.
• Comparing analyst concerns with executive responses through earnings divergence analysis.
• Tracking a brand across available graphs and generating brainstorm maps.
• Starting research and supplemental-context jobs, then checking their status.
• Consulting synthetic analysts and generating branded SVG data visualizations.

Partner-specific skills may add tools dynamically when enabled for an account. Account and personalization utilities may also be registered.

Limitations and notes

Access depends on Fodda account authentication and the graphs, partner skills, and subscription permissions available to that account. Search, synthesis, research, and visualization results are not all deterministic. The hosted service logs query text, graph access, and timestamps for billing, service improvement, and abuse prevention; it states that queries are not used to train AI models and that full AI-generated responses are not stored. The hosted privacy policy is authoritative for data handling.