Connect eDesigns to your AI assistant

eDesigns runs a public remote MCP server. Add it once and Claude, ChatGPT, Gemini CLI or your company's agents can evaluate your project and hand it to our team without leaving the conversation.

For marketing leaders, agencies and IT teams who already work inside an AI assistant. No account, API key or OAuth is needed: the server is read-only until the person explicitly agrees to be contacted.

Server endpoint

Model Context Protocol over Streamable HTTP. Paste this URL wherever your assistant asks for a remote MCP server or connector URL.

https://edesigns.com/api/mcp

Set-up by platform

Claude (web, desktop and mobile)

Custom connectors are available on Pro, Max, Team and Enterprise plans. Team and Enterprise admins can add it for the whole organization.

Open Settings → Connectors and choose “Add custom connector”.
Name it “eDesigns” and paste the MCP endpoint URL. Leave the OAuth fields empty.
Enable the connector in a chat (the attachments / tools menu) and ask Claude to evaluate your project with eDesigns.

ChatGPT (developer mode / Business, Enterprise, Edu)

Custom MCP connectors require developer mode (Settings → Apps & Connectors → Advanced) or a workspace admin on Business, Enterprise or Edu plans.

Go to Settings → Apps & Connectors → Create and choose an MCP server.
Name it “eDesigns”, paste the MCP endpoint URL and set authentication to “No authentication”.
Start a chat, enable the eDesigns connector under the tools menu and ask ChatGPT to evaluate your project with eDesigns.

Gemini CLI

Add the server to your project or user settings file (.gemini/settings.json).

Add the snippet below to ~/.gemini/settings.json (or the project's .gemini/settings.json).
Restart Gemini CLI and run /mcp to confirm that the eDesigns tools are listed.
Ask Gemini to evaluate your project with eDesigns; it will collect the nine answers and ask for your permission before sending them.

{
  "mcpServers": {
    "edesigns": {
      "httpUrl": "https://edesigns.com/api/mcp"
    }
  }
}

Other clients and corporate agents

Any MCP client that supports the Streamable HTTP transport (Cursor, Windsurf, VS Code, agent frameworks) can use the same URL.

Register the server as type “streamable-http” (sometimes just “http”) with the MCP endpoint URL.
No headers, tokens or environment variables are required.
Agents built on the Anthropic, OpenAI or Google SDKs can pass the URL directly as a remote MCP server.

What the assistant can do

Four tools, all built on the public API documented for developers. Read tools never store anything; write tools create a lead and only run after the person gives explicit consent.

get_company_profile (read-only) — Company facts, services, industries served and the published API product catalog.
get_evaluation_questions (read-only) — The nine qualification questions with their option keys and what each verdict means.
evaluate_project (creates a lead · consent required) — Submits the answers and the person's contact details; returns the verdict, a message to relay and, when qualified, a scheduling link. Requires explicit consent.
request_contact (creates a lead · consent required) — General inquiry or KYC / RENAP demo request on the person's behalf. Requires explicit consent.

Resources

llms.txt — The AI fact sheet of the website (company, services, FAQ, guidance for assistants).
openapi.json — The public OpenAPI specification the tools are built on.

Privacy and limits

The assistant must ask for your explicit permission before sending your name and email; without it the server refuses to create a lead.
Leads created through the server are marked as submitted by an AI agent, with the name of the assistant that connected, so our team knows how you arrived.
Requests are rate limited per network address, per session and per assistant; payloads are size-limited. The server keeps no session memory.
eDesigns publishes no email address; the server, the website forms and the public API are the official channels. Submissions are confidential and covered by the privacy policy.

For directories and platform teams

A machine-readable manifest in the MCP registry server.json format is published for connector directories. The server needs no authentication and follows the current MCP specification (Streamable HTTP, stateless). Server manifest (server.json)