@cognigy/mcp-server 0.1.0

package/QUICK_START.md

@@ -0,0 +1,195 @@
+# Quick Start Guide
+
+## What This Is
+
+A **Model Context Protocol (MCP) server** that lets AI assistants (Claude, Cursor) interact with Cognigy.AI through 9 high-level tools, including a **complete self-improvement loop** for AI Agents.
+
+## Key Feature: 100% Automatic AI Agent Setup
+
+**One tool call creates everything:**
+- ✅ AI Agent resource
+- ✅ Flow with AI Agent Job Node
+- ✅ REST Endpoint
+- ✅ **Agent is immediately testable!**
+
+## Installation (5 minutes)
+
+### 1. Install Dependencies
+
+```bash
+cd /path/to/cognigy/services/mcp-server
+npm install
+```
+
+### 2. Build
+
+```bash
+npm run build
+```
+
+### 3. Configure for Cursor
+
+Add to `~/Library/Application Support/Cursor/User/globalStorage/mcp/mcp.json`:
+
+```json
+{
+  "mcpServers": {
+    "cognigy": {
+      "command": "node",
+      "args": [
+        "/ABSOLUTE/PATH/TO/cognigy/services/mcp-server/dist/index.js"
+      ],
+      "env": {
+        "COGNIGY_API_BASE_URL": "https://api-trial.cognigy.ai",
+        "COGNIGY_API_KEY": "your-api-key-here"
+      }
+    }
+  }
+}
+```
+
+**Get your API key:**
+1. Login to Cognigy.AI
+2. User Menu → My Profile → API Keys
+3. Create new key, copy it
+
+### 4. Restart Cursor
+
+Settings → MCP → Toggle "cognigy" off/on (or restart Cursor completely)
+
+## Usage
+
+### Create an AI Agent (Fully Automatic!)
+
+In Cursor chat:
+```
+@cognigy Create an AI agent called "Support Bot" in project <project-id>
+```
+
+**This ONE command creates:**
+- AI Agent resource
+- Flow
+- AI Agent Job Node (connected to agent)
+- REST Endpoint
+- Returns endpoint URL
+
+### Talk to Your Agent
+
+```
+@cognigy Talk to the agent at <endpoint-url>
+Message: "I need help with something"
+```
+
+**Returns the agent's actual response!**
+
+### Improve Your Agent (The Loop!)
+
+```
+@cognigy Update agent <agent-id> description to be more empathetic and friendly
+```
+
+Then talk to it again and compare responses!
+
+### The Complete Self-Improvement Loop
+
+```
+1. @cognigy Create AI agent "Password Reset Helper" in project <id>
+   → Everything created, endpoint URL returned
+
+2. @cognigy Talk to agent at <url>: "I forgot my password"
+   → Agent responds (might be generic)
+
+3. Update agent description in Cognigy UI with better instructions
+   (Or use update tool if supported)
+
+4. @cognigy Talk to agent again: "I forgot my password"
+   → Compare responses - should be better!
+
+5. Repeat steps 3-4 until perfect!
+```
+
+## The 9 Tools
+
+1. **`manage_ai_agents`** ⭐ - Create/update AI agents (AUTO-CREATES EVERYTHING!)
+2. **`manage_knowledge`** - Knowledge stores for RAG
+3. **`manage_conversations`** - Query conversation history
+4. **`manage_flows`** - Create/modify flows
+5. **`manage_intents`** (Legacy) - Old NLU approach
+6. **`get_analytics`** - Metrics, logs, audit events
+7. **`manage_projects`** - Projects and endpoints
+8. **`manage_extensions`** - Custom functions
+9. **`talk_to_agent`** 🔄 - Talk to your agent, test responses
+
+## Important Notes
+
+### ID Formats (Critical!)
+
+Cognigy uses **two different ID formats**:
+
+**MongoDB _id** (24-char hex):
+- Format: `691b1c2e36bda2b14170fd73`
+- Used for: `projectId`, `aiAgentId`, most operations
+
+**UUID referenceId** (36-char with hyphens):
+- Format: `8184e6a8-e10c-4b03-8a83-538c004c321a`
+- Used for: `flowId` in endpoints
+
+When you create resources, you get both IDs. Use the right one for each operation!
+
+### System Prompt
+
+The MCP server includes a **comprehensive system prompt** as a resource that Claude/Cursor can read. This gives context about:
+- Cognigy.AI concepts (flows, agents, endpoints)
+- How tools work together
+- Best practices
+- Common workflows
+
+The AI assistant automatically becomes a Cognigy expert!
+
+### Modern Approach: AI Agents > Intents
+
+Focus on **LLM-based AI Agents** (tool #1), not traditional NLU/intents (tool #5 - legacy).
+
+## Troubleshooting
+
+### "COGNIGY_API_BASE_URL environment variable is required"
+→ Check your MCP config has the `env` section with API credentials
+
+### Agent not responding
+→ Make sure you have an LLM resource configured in the project (via Cognigy UI)
+
+### Invalid ID format errors
+→ Check if you're using the right ID format (MongoDB _id vs UUID referenceId)
+
+## What's Tested and Working
+
+✅ AI Agent creation with automatic flow/endpoint setup
+✅ REST endpoint creation
+✅ Talk to agent via endpoint
+✅ Agent responds properly
+✅ Update agent and see changes
+✅ Complete improvement loop
+
+## Next Steps
+
+1. **Get your project ID:**
+   ```
+   @cognigy List my projects
+   ```
+
+2. **Create your first agent:**
+   ```
+   @cognigy Create AI agent "My First Bot" in project <id>
+   ```
+
+3. **Start talking to it:**
+   ```
+   @cognigy Talk to agent at <endpoint-url>: "Hello!"
+   ```
+
+4. **Improve and iterate!** 🔄
+
+---
+
+**That's it!** You're ready to build and improve AI agents with the self-improvement loop! 🚀
+

package/README.md

@@ -0,0 +1,298 @@
+# Cognigy MCP Server
+
+> **📚 New here?** See **[INDEX.md](INDEX.md)** for quick navigation to the right documentation for your needs!
+
+A production-ready Model Context Protocol (MCP) server that provides AI assistants (Claude Desktop, Cursor IDE) with access to the Cognigy.AI REST API.
+
+## 🎯 Key Feature: Self-Improving AI Agents
+
+**Create a complete AI Agent setup in ONE tool call**, then iterate and improve through conversation:
+
+1. **Create** → AI Agent + Flow + Job Node + Endpoint (automatic!)
+2. **Test** → Talk to your agent via REST endpoint  
+3. **Improve** → Update job description based on responses
+4. **Test Again** → Compare responses and iterate
+5. **Repeat** → Until your agent is perfect!
+
+## Overview
+
+This MCP server exposes the Cognigy.AI API through 9 high-level workflow tools focused on **LLM-based AI Agents** (the modern approach), making it easy for AI assistants to build, test, and improve conversational AI systems.
+
+## Features
+
+- **9 High-Level Workflow Tools**: Grouped 359 API endpoints into 9 logical workflows (97.5% reduction!)
+- **100% Automatic AI Agent Setup**: One tool call creates Agent + Flow + Job Node + Endpoint
+- **Self-Improvement Loop**: Talk to agents, test responses, iterate until perfect
+- **System Prompt Included**: AI assistants automatically become Cognigy experts
+- **Authentication**: API Key authentication (configured via MCP)
+- **Rate Limiting**: Built-in protection against API abuse  
+- **Error Handling**: RFC 7807 compliant error responses
+- **Type Safety**: Full TypeScript with Zod validation
+- **Production Ready**: Logging, tests, multiple deployment options
+
+## Workflow Tools (AI Agent-Centric)
+
+1. **AI Agent Management** ⭐ - Create, update LLM-powered AI agents (PRIMARY TOOL)
+2. **Knowledge Management** - Manage knowledge stores for RAG-enabled AI agents
+3. **Conversation Management** - Query conversations to improve AI agent behavior
+4. **Flow Management** - Create flows with AI Agent Job Nodes
+5. **Intent & NLU Management** (Legacy) - Old approach, use AI Agents instead
+6. **Analytics & Monitoring** - Get metrics, logs, and audit events
+7. **Project & Endpoint Management** - Manage projects and REST endpoints
+8. **Extension Management** - Manage custom extensions and functions
+9. **Talk to AI Agent** 🔄 - **THE IMPROVEMENT LOOP** - Talk directly to your AI Agent via REST endpoint!
+
+**Important**: AI Agent job descriptions are configured in the **AI Agent Job Node** within a Flow, not in the Agent resource itself.
+
+**Modern Approach**: Focus on AI Agents (LLM-powered) rather than intents/NLU (legacy)
+
+### 🔄 The Self-Improvement Loop
+
+The **`talk_to_agent`** tool enables a powerful iterative improvement workflow:
+
+```
+┌─────────────────────────────────────────────────────┐
+│  1. Create/Update AI Agent (job description)       │
+└────────────────┬────────────────────────────────────┘
+                 ↓
+┌─────────────────────────────────────────────────────┐
+│  2. Talk to Agent (via REST endpoint)              │
+│     "How do I reset my password?"                  │
+└────────────────┬────────────────────────────────────┘
+                 ↓
+┌─────────────────────────────────────────────────────┐
+│  3. Evaluate Response                               │
+│     - Is it helpful?                               │
+│     - Is the tone right?                           │
+│     - Does it follow guidelines?                   │
+└────────────────┬────────────────────────────────────┘
+                 ↓
+┌─────────────────────────────────────────────────────┐
+│  4. Improve Job Description                        │
+│     - Add specific instructions                     │
+│     - Fix issues found                             │
+│     - Add edge case handling                       │
+└────────────────┬────────────────────────────────────┘
+                 ↓
+┌─────────────────────────────────────────────────────┐
+│  5. Talk to Agent Again (same question)            │
+│     Compare responses → Better? → Iterate          │
+└─────────────────────────────────────────────────────┘
+```
+
+**Human in the loop**: You provide feedback, suggest improvements, and approve changes!
+
+## Quick Start
+
+### Prerequisites
+
+- **Node.js 20+** (check with `node -v`)
+- **A Cognigy API key** — Cognigy.AI → User Menu → My Profile → API Keys → Create New
+
+### Add to your MCP client
+
+Add the following to your MCP client's config file. No build step, no cloning — `npx` handles everything automatically.
+
+```json
+{
+  "mcpServers": {
+    "cognigy": {
+      "command": "npx",
+      "args": ["-y", "@cognigy/mcp-server"],
+      "env": {
+        "COGNIGY_API_BASE_URL": "https://api-trial.cognigy.ai",
+        "COGNIGY_API_KEY": "your-api-key-here"
+      }
+    }
+  }
+}
+```
+
+Restart your MCP client, then try:
+```
+@cognigy List my Cognigy projects
+```
+
+**See [QUICK_START.md](QUICK_START.md) for detailed usage examples!**
+
+## Setup by MCP Client
+
+### Claude Desktop
+
+Edit `~/Library/Application Support/Claude/claude_desktop_config.json` and add the config block above. Restart Claude Desktop.
+
+### Cursor IDE
+
+Edit the MCP config file directly, or go to **Cursor > Settings > Cursor Settings > MCP > Add MCP Server**.
+
+**Config file location:**
+- macOS: `~/Library/Application Support/Cursor/User/globalStorage/mcp/mcp.json`
+- Windows: `%APPDATA%\Cursor\User\globalStorage\mcp\mcp.json`
+- Linux: `~/.config/Cursor/User/globalStorage/mcp/mcp.json`
+
+Add the config block above. Restart Cursor. Open the AI chat and type `@cognigy` to use the tools.
+
+### ChatGPT / Windsurf / Other MCP Clients
+
+Any MCP client that supports the **stdio** transport works. Add the same config block to your client's MCP settings.
+
+## Configuration
+
+All configuration is passed via the `env` block in the MCP config. No `.env` file needed.
+
+| Variable | Required | Default | Description |
+|---|---|---|---|
+| `COGNIGY_API_BASE_URL` | Yes | — | Your Cognigy API URL |
+| `COGNIGY_API_KEY` | Yes | — | Your Cognigy API key |
+| `LOG_LEVEL` | No | `info` | `debug`, `info`, `warn`, `error` |
+| `RATE_LIMIT_MAX_REQUESTS` | No | `100` | Max requests per window |
+| `RATE_LIMIT_WINDOW_MS` | No | `60000` | Rate limit window (ms) |
+
+### Getting Your API Key
+
+1. Log in to Cognigy.AI
+2. **User Menu → My Profile**
+3. **API Keys** section
+4. Click **Create API Key**
+5. Copy the key
+
+<details>
+<summary>Local development setup (for contributors)</summary>
+
+If you're developing the MCP server itself:
+
+```bash
+git clone <repo-url>
+cd cognigy-mcp
+npm install
+npm run build
+```
+
+Then point your MCP client to the local build:
+
+```json
+{
+  "mcpServers": {
+    "cognigy": {
+      "command": "node",
+      "args": ["/absolute/path/to/cognigy-mcp/dist/index.js"],
+      "env": {
+        "COGNIGY_API_BASE_URL": "https://api-trial.cognigy.ai",
+        "COGNIGY_API_KEY": "your-api-key-here"
+      }
+    }
+  }
+}
+```
+
+Or use the interactive setup: `npm run setup`
+
+</details>
+
+### System Prompt (Included!)
+
+The MCP server includes a comprehensive **system prompt as an MCP resource** that Claude automatically loads. This transforms Claude into a Cognigy expert that:
+
+- ✅ Understands modern **LLM-based AI Agent architecture**
+- ✅ Focuses on **creating and refining AI Agents** (not legacy NLU/intents)
+- ✅ Knows the AI Agent workflow: Agent resource → LLM resource → AI Agent Job Node
+- ✅ Plans multi-step workflows automatically
+- ✅ Suggests best practices for AI Agent job descriptions
+- ✅ Guides users to iterate and improve agent behavior
+
+**No additional configuration needed** - it works automatically! See `docs/SYSTEM_PROMPT.md` for details.
+
+## Architecture
+
+The MCP server follows a modular architecture:
+
+- `src/index.ts` - Main server initialization and MCP protocol handling
+- `src/tools/` - Tool handlers for each workflow category
+- `src/api/` - API client and request handlers
+- `src/schemas/` - Zod schemas for input validation
+- `src/utils/` - Utilities (logging, rate limiting, etc.)
+- `src/resources/` - MCP resources (documentation, examples)
+
+## Security
+
+- API keys are never logged or exposed
+- All inputs are validated using Zod schemas
+- Rate limiting prevents API abuse
+- Supports OAuth2 for enterprise deployments
+
+## Project Statistics
+
+- **9 Tools**: Instead of 359 endpoint-specific tools (97.5% reduction)
+- **34 Operations**: Covering all major Cognigy workflows + agent testing
+- **~115 API Endpoints**: Accessible through workflow grouping
+- **Full Type Safety**: TypeScript with Zod validation
+- **Production Ready**: Rate limiting, logging, error handling, tests
+- **🔄 100% Automated**: Create AI Agent → Flow → Job Node → Endpoint in one call
+- **✅ Tested & Working**: Complete self-improvement loop validated end-to-end
+
+## What's Working (Verified!)
+
+✅ **Automatic AI Agent Setup** - Creates agent, flow, job node, and endpoint in one tool call
+✅ **Talk to Agent** - Sends messages and receives responses via REST endpoint
+✅ **Agent Updates** - Modify agent configuration and see changes immediately
+✅ **Complete Improvement Loop** - Create → Test → Improve → Test → Iterate
+✅ **All Infrastructure Tools** - Projects, flows, endpoints, knowledge, analytics
+✅ **System Prompt** - AI assistants have full Cognigy domain expertise
+
+## Important ID Format Note
+
+Cognigy uses two ID formats:
+- **MongoDB _id** (24-char hex): `691b1c2e36bda2b14170fd73` - Used for most operations
+- **UUID referenceId** (36-char): `8184e6a8-e10c-4b03-8a83-538c004c321a` - Used for flowId in endpoints
+
+When creating resources, both IDs are returned. The tools handle this automatically.
+
+## Key Design Decision: Why Only 9 Tools?
+
+The Cognigy API has ~359 endpoints across 50+ categories. Creating one tool per endpoint would overwhelm AI agents and make the system unusable. Our solution:
+
+**Workflow-Based Grouping**: Group related endpoints into high-level workflows that represent what users actually want to accomplish:
+
+- ✅ Natural for AI agents to understand
+- ✅ Matches human mental models
+- ✅ Easy to discover and use
+- ✅ Maintainable and extensible
+
+Example: Instead of separate tools for `createAgent`, `getAgent`, `updateAgent`, `deleteAgent`, `listAgents`, `hireAgent`, we have one `manage_ai_agents` tool with an `operation` parameter.
+
+## Modern AI Agent-Centric Approach
+
+The MCP server focuses on **LLM-based AI Agents** (the modern approach) rather than traditional NLU/intents (legacy).
+
+**What Happens When You Create an AI Agent:**
+1. ✅ AI Agent resource created
+2. ✅ Flow created automatically
+3. ✅ AI Agent Job Node added to flow (connected to agent)
+4. ✅ REST Endpoint created (ready to use!)
+5. ✅ Endpoint URL returned
+
+**Then:** Use `talk_to_agent` tool to test responses and iterate!
+
+**Note:** You'll need an LLM resource (GPT-4, Claude, etc.) configured in the project via Cognigy UI for the agent to work.
+
+## Documentation Navigation
+
+- **Quick Start**: See [QUICK_START.md](QUICK_START.md) for setup & usage examples
+- **Technical Details**: See [PROJECT_SUMMARY.md](PROJECT_SUMMARY.md) for architecture
+- **For Developers**: See [HANDOFF.md](HANDOFF.md) for implementation details
+- **All Docs**: See [INDEX.md](INDEX.md) for complete documentation index
+
+## Contributing
+
+Contributions are welcome! Read [.cursorrules](.cursorrules) for development guidelines and [HANDOFF.md](HANDOFF.md) for technical details.
+
+## Support
+
+- **Documentation**: See `docs/` directory
+- **Issues**: GitHub issues
+- **Cognigy API Support**: support@cognigy.com
+
+## License
+
+MIT
+

package/dist/api/client.d.ts

@@ -0,0 +1,38 @@
+/**
+ * Cognigy API client for making HTTP requests
+ */
+import { AxiosRequestConfig } from 'axios';
+export interface CognigyApiClientConfig {
+    baseUrl: string;
+    apiKey: string;
+}
+export declare class CognigyApiClient {
+    private client;
+    private apiKey;
+    constructor(config: CognigyApiClientConfig);
+    /**
+     * Format error responses according to RFC 7807
+     */
+    private formatError;
+    /**
+     * Make a GET request
+     */
+    get<T = any>(url: string, config?: AxiosRequestConfig): Promise<T>;
+    /**
+     * Make a POST request
+     */
+    post<T = any>(url: string, data?: any, config?: AxiosRequestConfig): Promise<T>;
+    /**
+     * Make a PUT request
+     */
+    put<T = any>(url: string, data?: any, config?: AxiosRequestConfig): Promise<T>;
+    /**
+     * Make a PATCH request
+     */
+    patch<T = any>(url: string, data?: any, config?: AxiosRequestConfig): Promise<T>;
+    /**
+     * Make a DELETE request
+     */
+    delete<T = any>(url: string, config?: AxiosRequestConfig): Promise<T>;
+}
+//# sourceMappingURL=client.d.ts.map

package/dist/api/client.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"client.d.ts","sourceRoot":"","sources":["../../src/api/client.ts"],"names":[],"mappings":"AAAA;;GAEG;AAEH,OAAc,EAAiB,kBAAkB,EAAiB,MAAM,OAAO,CAAC;AAGhF,MAAM,WAAW,sBAAsB;IACrC,OAAO,EAAE,MAAM,CAAC;IAChB,MAAM,EAAE,MAAM,CAAC;CAChB;AAED,qBAAa,gBAAgB;IAC3B,OAAO,CAAC,MAAM,CAAgB;IAC9B,OAAO,CAAC,MAAM,CAAS;gBAEX,MAAM,EAAE,sBAAsB;IA+C1C;;OAEG;IACH,OAAO,CAAC,WAAW;IAcnB;;OAEG;IACG,GAAG,CAAC,CAAC,GAAG,GAAG,EAAE,GAAG,EAAE,MAAM,EAAE,MAAM,CAAC,EAAE,kBAAkB,GAAG,OAAO,CAAC,CAAC,CAAC;IAKxE;;OAEG;IACG,IAAI,CAAC,CAAC,GAAG,GAAG,EAAE,GAAG,EAAE,MAAM,EAAE,IAAI,CAAC,EAAE,GAAG,EAAE,MAAM,CAAC,EAAE,kBAAkB,GAAG,OAAO,CAAC,CAAC,CAAC;IAKrF;;OAEG;IACG,GAAG,CAAC,CAAC,GAAG,GAAG,EAAE,GAAG,EAAE,MAAM,EAAE,IAAI,CAAC,EAAE,GAAG,EAAE,MAAM,CAAC,EAAE,kBAAkB,GAAG,OAAO,CAAC,CAAC,CAAC;IAKpF;;OAEG;IACG,KAAK,CAAC,CAAC,GAAG,GAAG,EAAE,GAAG,EAAE,MAAM,EAAE,IAAI,CAAC,EAAE,GAAG,EAAE,MAAM,CAAC,EAAE,kBAAkB,GAAG,OAAO,CAAC,CAAC,CAAC;IAKtF;;OAEG;IACG,MAAM,CAAC,CAAC,GAAG,GAAG,EAAE,GAAG,EAAE,MAAM,EAAE,MAAM,CAAC,EAAE,kBAAkB,GAAG,OAAO,CAAC,CAAC,CAAC;CAI5E"}

package/dist/api/client.js

@@ -0,0 +1,97 @@
+/**
+ * Cognigy API client for making HTTP requests
+ */
+import axios from 'axios';
+import { logger } from '../utils/logger.js';
+export class CognigyApiClient {
+    client;
+    apiKey;
+    constructor(config) {
+        this.apiKey = config.apiKey;
+        this.client = axios.create({
+            baseURL: config.baseUrl,
+            headers: {
+                'Content-Type': 'application/json',
+                'Accept': 'application/json',
+            },
+            timeout: 30000,
+        });
+        // Add request interceptor for authentication
+        this.client.interceptors.request.use((config) => {
+            config.headers['X-API-Key'] = this.apiKey;
+            logger.debug(`API Request: ${config.method?.toUpperCase()} ${config.url}`);
+            return config;
+        }, (error) => {
+            logger.error('API Request Error', { error: error.message });
+            return Promise.reject(error);
+        });
+        // Add response interceptor for logging
+        this.client.interceptors.response.use((response) => {
+            logger.debug(`API Response: ${response.status} ${response.config.url}`);
+            return response;
+        }, (error) => {
+            const status = error.response?.status || 'N/A';
+            const message = error.response?.data?.detail || error.message;
+            const traceId = error.response?.data?.traceId;
+            logger.error('API Response Error', {
+                status,
+                message,
+                traceId,
+                url: error.config?.url,
+            });
+            return Promise.reject(this.formatError(error));
+        });
+    }
+    /**
+     * Format error responses according to RFC 7807
+     */
+    formatError(error) {
+        if (error.response?.data) {
+            const data = error.response.data;
+            const message = data.detail || data.title || 'API request failed';
+            const enhancedError = new Error(message);
+            enhancedError.status = data.status || error.response.status;
+            enhancedError.code = data.code;
+            enhancedError.traceId = data.traceId;
+            enhancedError.details = data.details;
+            return enhancedError;
+        }
+        return error;
+    }
+    /**
+     * Make a GET request
+     */
+    async get(url, config) {
+        const response = await this.client.get(url, config);
+        return response.data;
+    }
+    /**
+     * Make a POST request
+     */
+    async post(url, data, config) {
+        const response = await this.client.post(url, data, config);
+        return response.data;
+    }
+    /**
+     * Make a PUT request
+     */
+    async put(url, data, config) {
+        const response = await this.client.put(url, data, config);
+        return response.data;
+    }
+    /**
+     * Make a PATCH request
+     */
+    async patch(url, data, config) {
+        const response = await this.client.patch(url, data, config);
+        return response.data;
+    }
+    /**
+     * Make a DELETE request
+     */
+    async delete(url, config) {
+        const response = await this.client.delete(url, config);
+        return response.data;
+    }
+}
+//# sourceMappingURL=client.js.map

package/dist/api/client.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"client.js","sourceRoot":"","sources":["../../src/api/client.ts"],"names":[],"mappings":"AAAA;;GAEG;AAEH,OAAO,KAA2D,MAAM,OAAO,CAAC;AAChF,OAAO,EAAE,MAAM,EAAE,MAAM,oBAAoB,CAAC;AAO5C,MAAM,OAAO,gBAAgB;IACnB,MAAM,CAAgB;IACtB,MAAM,CAAS;IAEvB,YAAY,MAA8B;QACxC,IAAI,CAAC,MAAM,GAAG,MAAM,CAAC,MAAM,CAAC;QAC5B,IAAI,CAAC,MAAM,GAAG,KAAK,CAAC,MAAM,CAAC;YACzB,OAAO,EAAE,MAAM,CAAC,OAAO;YACvB,OAAO,EAAE;gBACP,cAAc,EAAE,kBAAkB;gBAClC,QAAQ,EAAE,kBAAkB;aAC7B;YACD,OAAO,EAAE,KAAK;SACf,CAAC,CAAC;QAEH,6CAA6C;QAC7C,IAAI,CAAC,MAAM,CAAC,YAAY,CAAC,OAAO,CAAC,GAAG,CAClC,CAAC,MAAM,EAAE,EAAE;YACT,MAAM,CAAC,OAAO,CAAC,WAAW,CAAC,GAAG,IAAI,CAAC,MAAM,CAAC;YAC1C,MAAM,CAAC,KAAK,CAAC,gBAAgB,MAAM,CAAC,MAAM,EAAE,WAAW,EAAE,IAAI,MAAM,CAAC,GAAG,EAAE,CAAC,CAAC;YAC3E,OAAO,MAAM,CAAC;QAChB,CAAC,EACD,CAAC,KAAK,EAAE,EAAE;YACR,MAAM,CAAC,KAAK,CAAC,mBAAmB,EAAE,EAAE,KAAK,EAAE,KAAK,CAAC,OAAO,EAAE,CAAC,CAAC;YAC5D,OAAO,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;QAC/B,CAAC,CACF,CAAC;QAEF,uCAAuC;QACvC,IAAI,CAAC,MAAM,CAAC,YAAY,CAAC,QAAQ,CAAC,GAAG,CACnC,CAAC,QAAQ,EAAE,EAAE;YACX,MAAM,CAAC,KAAK,CAAC,iBAAiB,QAAQ,CAAC,MAAM,IAAI,QAAQ,CAAC,MAAM,CAAC,GAAG,EAAE,CAAC,CAAC;YACxE,OAAO,QAAQ,CAAC;QAClB,CAAC,EACD,CAAC,KAAK,EAAE,EAAE;YACR,MAAM,MAAM,GAAG,KAAK,CAAC,QAAQ,EAAE,MAAM,IAAI,KAAK,CAAC;YAC/C,MAAM,OAAO,GAAG,KAAK,CAAC,QAAQ,EAAE,IAAI,EAAE,MAAM,IAAI,KAAK,CAAC,OAAO,CAAC;YAC9D,MAAM,OAAO,GAAG,KAAK,CAAC,QAAQ,EAAE,IAAI,EAAE,OAAO,CAAC;YAE9C,MAAM,CAAC,KAAK,CAAC,oBAAoB,EAAE;gBACjC,MAAM;gBACN,OAAO;gBACP,OAAO;gBACP,GAAG,EAAE,KAAK,CAAC,MAAM,EAAE,GAAG;aACvB,CAAC,CAAC;YAEH,OAAO,OAAO,CAAC,MAAM,CAAC,IAAI,CAAC,WAAW,CAAC,KAAK,CAAC,CAAC,CAAC;QACjD,CAAC,CACF,CAAC;IACJ,CAAC;IAED;;OAEG;IACK,WAAW,CAAC,KAAU;QAC5B,IAAI,KAAK,CAAC,QAAQ,EAAE,IAAI,EAAE,CAAC;YACzB,MAAM,IAAI,GAAG,KAAK,CAAC,QAAQ,CAAC,IAAI,CAAC;YACjC,MAAM,OAAO,GAAG,IAAI,CAAC,MAAM,IAAI,IAAI,CAAC,KAAK,IAAI,oBAAoB,CAAC;YAClE,MAAM,aAAa,GAAG,IAAI,KAAK,CAAC,OAAO,CAAC,CAAC;YACxC,aAAqB,CAAC,MAAM,GAAG,IAAI,CAAC,MAAM,IAAI,KAAK,CAAC,QAAQ,CAAC,MAAM,CAAC;YACpE,aAAqB,CAAC,IAAI,GAAG,IAAI,CAAC,IAAI,CAAC;YACvC,aAAqB,CAAC,OAAO,GAAG,IAAI,CAAC,OAAO,CAAC;YAC7C,aAAqB,CAAC,OAAO,GAAG,IAAI,CAAC,OAAO,CAAC;YAC9C,OAAO,aAAa,CAAC;QACvB,CAAC;QACD,OAAO,KAAK,CAAC;IACf,CAAC;IAED;;OAEG;IACH,KAAK,CAAC,GAAG,CAAU,GAAW,EAAE,MAA2B;QACzD,MAAM,QAAQ,GAAqB,MAAM,IAAI,CAAC,MAAM,CAAC,GAAG,CAAC,GAAG,EAAE,MAAM,CAAC,CAAC;QACtE,OAAO,QAAQ,CAAC,IAAI,CAAC;IACvB,CAAC;IAED;;OAEG;IACH,KAAK,CAAC,IAAI,CAAU,GAAW,EAAE,IAAU,EAAE,MAA2B;QACtE,MAAM,QAAQ,GAAqB,MAAM,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,EAAE,IAAI,EAAE,MAAM,CAAC,CAAC;QAC7E,OAAO,QAAQ,CAAC,IAAI,CAAC;IACvB,CAAC;IAED;;OAEG;IACH,KAAK,CAAC,GAAG,CAAU,GAAW,EAAE,IAAU,EAAE,MAA2B;QACrE,MAAM,QAAQ,GAAqB,MAAM,IAAI,CAAC,MAAM,CAAC,GAAG,CAAC,GAAG,EAAE,IAAI,EAAE,MAAM,CAAC,CAAC;QAC5E,OAAO,QAAQ,CAAC,IAAI,CAAC;IACvB,CAAC;IAED;;OAEG;IACH,KAAK,CAAC,KAAK,CAAU,GAAW,EAAE,IAAU,EAAE,MAA2B;QACvE,MAAM,QAAQ,GAAqB,MAAM,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,GAAG,EAAE,IAAI,EAAE,MAAM,CAAC,CAAC;QAC9E,OAAO,QAAQ,CAAC,IAAI,CAAC;IACvB,CAAC;IAED;;OAEG;IACH,KAAK,CAAC,MAAM,CAAU,GAAW,EAAE,MAA2B;QAC5D,MAAM,QAAQ,GAAqB,MAAM,IAAI,CAAC,MAAM,CAAC,MAAM,CAAC,GAAG,EAAE,MAAM,CAAC,CAAC;QACzE,OAAO,QAAQ,CAAC,IAAI,CAAC;IACvB,CAAC;CACF"}

package/dist/config.d.ts

@@ -0,0 +1,19 @@
+/**
+ * Configuration for the Cognigy MCP Server
+ */
+export interface Config {
+    apiBaseUrl: string;
+    apiKey: string;
+    serverName: string;
+    serverVersion: string;
+    logLevel: 'debug' | 'info' | 'warn' | 'error';
+    rateLimit: {
+        maxRequests: number;
+        windowMs: number;
+    };
+}
+/**
+ * Load configuration from environment variables
+ */
+export declare function loadConfig(): Config;
+//# sourceMappingURL=config.d.ts.map

package/dist/config.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"config.d.ts","sourceRoot":"","sources":["../src/config.ts"],"names":[],"mappings":"AAAA;;GAEG;AACH,MAAM,WAAW,MAAM;IACrB,UAAU,EAAE,MAAM,CAAC;IACnB,MAAM,EAAE,MAAM,CAAC;IACf,UAAU,EAAE,MAAM,CAAC;IACnB,aAAa,EAAE,MAAM,CAAC;IACtB,QAAQ,EAAE,OAAO,GAAG,MAAM,GAAG,MAAM,GAAG,OAAO,CAAC;IAC9C,SAAS,EAAE;QACT,WAAW,EAAE,MAAM,CAAC;QACpB,QAAQ,EAAE,MAAM,CAAC;KAClB,CAAC;CACH;AAED;;GAEG;AACH,wBAAgB,UAAU,IAAI,MAAM,CAuBnC"}

package/dist/config.js

@@ -0,0 +1,25 @@
+/**
+ * Load configuration from environment variables
+ */
+export function loadConfig() {
+    const apiBaseUrl = process.env.COGNIGY_API_BASE_URL;
+    const apiKey = process.env.COGNIGY_API_KEY;
+    if (!apiBaseUrl) {
+        throw new Error('COGNIGY_API_BASE_URL environment variable is required');
+    }
+    if (!apiKey) {
+        throw new Error('COGNIGY_API_KEY environment variable is required');
+    }
+    return {
+        apiBaseUrl,
+        apiKey,
+        serverName: process.env.MCP_SERVER_NAME || 'cognigy-api-mcp',
+        serverVersion: process.env.MCP_SERVER_VERSION || '1.0.0',
+        logLevel: process.env.LOG_LEVEL || 'info',
+        rateLimit: {
+            maxRequests: parseInt(process.env.RATE_LIMIT_MAX_REQUESTS || '100', 10),
+            windowMs: parseInt(process.env.RATE_LIMIT_WINDOW_MS || '60000', 10),
+        },
+    };
+}
+//# sourceMappingURL=config.js.map