Live MCP Server

Install a real AgentFaktory MCP demo server.

Connect ChatGPT, Claude, or any MCP-capable client to a public read-only tool server. The best tools are AI-enhanced on AgentFaktory's server: they build board-pack style roadmaps, answer AgentFaktory website questions, simulate governed agent runs with concrete records, estimate ROI, and prepare discovery briefs.

Install MCP serverView MCP solution
Remote MCP endpoint
10 demo tools
Read-only and safe
Useful business outputs
Try these prompts
Using the AgentFaktory MCP server, build an AI board pack for a logistics company trying to reduce manual dispatch coordination and improve margin visibility.
Ask the AgentFaktory MCP server who is behind the company, what services it offers, and which solution best fits an enterprise finance automation buyer.
Simulate a governed AI agent run for invoice exception handling across SAP, SharePoint, email approvals, and a finance data warehouse.
Design an agent architecture for invoice exception handling across SAP, SharePoint, and email approvals.
Estimate ROI for a 12-person operations team spending 6 hours per person per week reconciling spreadsheets.
Demo tools
answer_agentfaktory_question

Answers questions about AgentFaktory, leadership, services, solutions, industries, approach, and contact options.

get_agentfaktory_profile

Returns a full AgentFaktory profile with team, positioning, services, solutions, industries, and differentiation.

build_ai_board_pack

Creates an executive AI transformation plan with proof demo scenario, roadmap, risk register, first pilots, and decision ask.

simulate_agent_workflow

Simulates a governed AI agent run with sample records, concrete exceptions, tool-call trace, evidence, approvals, and action packet.

map_ai_opportunities

Ranks practical AI opportunities for an industry, challenge, systems, and constraints.

design_agent_architecture

Turns a workflow into a safe agent/MCP architecture with guardrails and delivery plan.

estimate_roi

Builds a directional automation ROI scenario from team size, effort, and cost assumptions.

generate_discovery_brief

Creates a discovery-call brief with pains, questions, and relevant AgentFaktory services.

search_agentfaktory_services

Searches AgentFaktory services and solutions by problem, industry, or keyword.

request_agentfaktory_followup

Creates a lead after the MCP client asks and the user explicitly agrees to AgentFaktory follow-up.

Remote MCP URLhttps://agentfaktory.ai/api/mcp
Open ChatGPT settings

In ChatGPT, add a custom app/connector with the URL above. Exact availability depends on the ChatGPT plan, workspace settings, and developer/app access.

Claude Desktop

Download the desktop extension bundle and open it with Claude Desktop where supported.

Download .mcpbClaude install guide

Claude API config

For API-based testing, copy the server definition and add it to your MCP server list.

{
  "type": "url",
  "url": "https://agentfaktory.ai/api/mcp",
  "name": "agentfaktory-demo"
}
Need help installing?

Send me MCP setup help

Privacy and lead handling

The demo MCP endpoint logs minimal operational analytics such as tool name, timestamp, user agent, and an IP hash. It does not require authentication and does not write to your systems. Planning tools may be processed with AgentFaktory's server-side AI provider to generate richer outputs. MCP clients are instructed to ask whether you want AgentFaktory follow-up after useful outputs; contact details are only submitted if you agree. The install-help form is treated as a lead request and follows the Privacy Policy.