# VikrantSingh01/adaptive-cards-mcp [Health: Active]

**Category:** 💬 Communication  
**Repository:** https://github.com/VikrantSingh01/adaptive-cards-mcp  
**GitHub Stars:** 33  
**npm Downloads (last month):** 204  
**Views:** 3  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/vikrantsingh01-adaptive-cards-mcp

## Description
AI-powered Adaptive Card generation for Teams, Outlook, Copilot, and ChatGPT. 9 MCP tools for generating, validating, optimizing, templating, and transforming cards with accessibility scoring and host compatibility checks.

## Tools
Capabilities this server exposes over MCP:

- **generate_card** — Convert any content — natural language description, structured data, or a combination — into a valid Adaptive Card v1.6 JSON. Returns cardId for reference in subsequent tool calls.
- **validate_card** — Validate an Adaptive Card JSON against the v1.6 schema. Returns diagnostics with suggested fixes for each error. Accepts card JSON or a cardId from a previous tool call.
- **data_to_card** — Convert structured data (JSON array, CSV, key-value object) into the optimal Adaptive Card presentation.
- **optimize_card** — Optimize an existing Adaptive Card. Accepts card JSON or a cardId.
- **template_card** — Convert a static Adaptive Card into an Adaptive Card Template with ${expression} data binding.
- **transform_card** — Transform an Adaptive Card: upgrade/downgrade version, apply host-specific constraints, or flatten nesting.
- **suggest_layout** — Recommend the best Adaptive Card layout pattern for a given description.
- **generate_and_validate** — Generate an Adaptive Card and immediately validate + optionally optimize it in a single call. Reduces tool-call overhead for common workflows.
- **card_workflow** — Execute a multi-step card pipeline in a single call. Steps: generate, validate, optimize, template, transform.

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `npx` (confidence: high):

```json
"mcpServers": {
  "adaptive-cards-mcp": {
    "command": "npx",
    "args": ["-y","adaptive-cards-mcp"],
    "env": {
      "TRANSPORT": "",
      "PORT": "",
      "MCP_API_KEY": ""
    }
  }
}
```

**Requires environment variables:** `TRANSPORT`, `PORT`, `MCP_API_KEY` — the values above are empty placeholders; fill in real credentials before running (see the repository for what each one is for).

## Documentation

## What VikrantSingh01/adaptive-cards-mcp MCP server does

VikrantSingh01/adaptive-cards-mcp MCP server gives AI assistants a toolset for producing and modifying Adaptive Card JSON. It targets Adaptive Cards version 1.6 and supports workflows such as approval forms, notifications, reports, dashboards, profiles, and other card-based interfaces. Input can be a plain-language description, structured JSON, CSV, key-value data, or an existing card.

The output is intended for use with surfaces such as Microsoft Teams, Outlook, Copilot, and ChatGPT. Host and version differences can be handled during transformation or validation, including requests to make a card suitable for an older Adaptive Cards version or a particular host.

## How it works

A typical request starts with generation from a description or data. The server can return a card identifier, allowing later calls to refer to that card instead of sending the complete JSON again. Existing card JSON can also be passed directly to validation, optimization, templating, or transformation tools.

Validation checks the card against the v1.6 schema and reports diagnostics with suggested corrections. Optimization can address accessibility and presentation concerns, while host checks help identify compatibility issues. A combined operation can generate, validate, and optionally optimize a card in one call. The workflow operation runs a selected sequence of generation, validation, optimization, templating, and transformation steps.

## Setup and configuration

The package can be launched without a local installation by using npx. For a local stdio connection, configure an MCP client to run `npx adaptive-cards-mcp`.

The README provides configurations for Claude Code, GitHub Copilot in VS Code, Cursor, and Windsurf. For remote connections, run the server with `TRANSPORT=sse` and set a port such as `PORT=3001`. An optional `MCP_API_KEY` can be supplied to enable authentication for the SSE deployment.

Copilot Studio can connect to the server through its Model Context Protocol tool configuration and an HTTPS server URL ending in `/sse`. ChatGPT requires Developer mode, a connector configuration, and an HTTPS MCP URL. The repository does not describe additional environment variables or a separate build requirement for the npx workflow.

## Tools and capabilities

VikrantSingh01/adaptive-cards-mcp exposes nine MCP tools:

- `generate_card` creates v1.6 card JSON from descriptions, structured data, or both.
- `validate_card` checks schema validity and returns diagnostics and suggested fixes.
- `data_to_card` chooses a card presentation for arrays, CSV, or key-value objects.
- `optimize_card` improves an existing card, including accessibility-related details.
- `template_card` replaces static values with `${expression}` data bindings.
- `transform_card` changes versions, applies host constraints, or flattens nesting.
- `suggest_layout` recommends a layout pattern with a rationale.
- `generate_and_validate` combines generation with validation and optional optimization.
- `card_workflow` executes a multi-step card pipeline in one request.

The server also provides guided MCP prompts for creating a card, converting data, and reviewing an existing card. These prompts describe sequences such as generate, validate, optimize, and host configuration, or validate, accessibility-check, auto-fix, and report.

## Limitations and notes

Generated cards still need to be checked against the actual host where they will render. Adaptive Card feature support varies by host and version, so a card that works in Teams may need transformation for Outlook, Webex, or another target. The material documents SSE authentication through `MCP_API_KEY`, but it does not describe a broader authorization model.

_Full upstream README: https://allmcps.com/mcp/vikrantsingh01-adaptive-cards-mcp/readme_