|
| 1 | +--- |
| 2 | +title: Test MCP servers with playground |
| 3 | +description: |
| 4 | + Use the playground feature to test and validate MCP servers directly in the |
| 5 | + ToolHive UI with AI model providers. |
| 6 | +--- |
| 7 | + |
| 8 | +import useBaseUrl from '@docusaurus/useBaseUrl'; |
| 9 | +import ThemedImage from '@theme/ThemedImage'; |
| 10 | + |
| 11 | +ToolHive's playground lets you test and validate MCP servers directly in the UI |
| 12 | +without requiring additional client setup. This streamlined testing environment |
| 13 | +helps you quickly evaluate functionality and behavior before deploying MCP |
| 14 | +servers to production environments. |
| 15 | + |
| 16 | +## Key capabilities |
| 17 | + |
| 18 | +### Instant testing of MCP servers |
| 19 | + |
| 20 | +Enter an AI model API key, select your MCP servers and tools, and begin testing |
| 21 | +immediately in the desktop app. The playground eliminates the friction of |
| 22 | +setting up external AI clients just to validate that your MCP servers work |
| 23 | +correctly. |
| 24 | + |
| 25 | +### Detailed interaction logs |
| 26 | + |
| 27 | +See tool details, parameters, and execution results directly in the UI, ensuring |
| 28 | +full visibility into tool performance and responses. Every interaction is |
| 29 | +logged, making it easy to understand exactly what your MCP servers are doing and |
| 30 | +how they respond to requests. |
| 31 | + |
| 32 | +### Integrated ToolHive management |
| 33 | + |
| 34 | +The playground includes a built-in MCP server that lets you manage your other |
| 35 | +MCP servers directly through natural language commands. You can list servers, |
| 36 | +check their status, start or stop them, and perform other management tasks using |
| 37 | +conversational AI. |
| 38 | + |
| 39 | +## Getting started |
| 40 | + |
| 41 | +To start using the playground: |
| 42 | + |
| 43 | +1. **Access the playground**: Click the **Playground** tab in the ToolHive UI |
| 44 | + navigation bar. |
| 45 | + |
| 46 | +2. **Configure API keys**: Click **Configure your API Keys** to set up access to |
| 47 | + AI model providers: |
| 48 | + - **OpenAI**: Enter your OpenAI API key to use GPT models |
| 49 | + - **Anthropic**: Add your Anthropic API key for Claude models |
| 50 | + - **Google**: Configure Google AI API key for Gemini models |
| 51 | + - **xAI**: Set up xAI API key for Grok models |
| 52 | + - **OpenRouter**: Add OpenRouter API key for access to multiple model |
| 53 | + providers |
| 54 | + |
| 55 | +3. **Select MCP tools**: Click the tools icon to manage which MCP servers and |
| 56 | + tools are available in the playground. |
| 57 | + |
| 58 | +<ThemedImage |
| 59 | + alt='ToolHive playground tools management showing available MCP tools' |
| 60 | + sources={{ |
| 61 | + light: useBaseUrl( |
| 62 | + '/img/toolhive/toolhive-ui-playground-thv-mcp-light.webp' |
| 63 | + ), |
| 64 | + dark: useBaseUrl('/img/toolhive/toolhive-ui-playground-thv-mcp-dark.webp'), |
| 65 | + }} |
| 66 | + title='MCP tools management interface' |
| 67 | +/> |
| 68 | + |
| 69 | +- View all your running MCP servers |
| 70 | +- Enable or disable specific tools from each server |
| 71 | +- Search and filter tools by name or functionality |
| 72 | +- The `toolhive mcp` server is included by default, providing management |
| 73 | + capabilities |
| 74 | + |
| 75 | +4. **Start testing**: Begin chatting with your chosen AI model. The model will |
| 76 | + have access to all enabled MCP tools and can execute them based on your |
| 77 | + requests. |
| 78 | + |
| 79 | +## Using the playground |
| 80 | + |
| 81 | +### Testing MCP server functionality |
| 82 | + |
| 83 | +Use the playground to validate that your MCP servers work as expected: |
| 84 | + |
| 85 | +```text |
| 86 | +Can you list all my MCP servers and show their current status? |
| 87 | +``` |
| 88 | + |
| 89 | +The AI will use the `list_servers` tool from the ToolHive MCP server to provide |
| 90 | +a comprehensive overview of your server status. |
| 91 | + |
| 92 | +<ThemedImage |
| 93 | + alt='ToolHive playground showing AI response with MCP tool execution results' |
| 94 | + sources={{ |
| 95 | + light: useBaseUrl( |
| 96 | + '/img/toolhive/toolhive-ui-playground-chat-response-light.webp' |
| 97 | + ), |
| 98 | + dark: useBaseUrl( |
| 99 | + '/img/toolhive/toolhive-ui-playground-chat-response-dark.webp' |
| 100 | + ), |
| 101 | + }} |
| 102 | + title='Chat response showing tool execution and results' |
| 103 | +/> |
| 104 | + |
| 105 | +Or test that an individual MCP tool is working as expected: |
| 106 | + |
| 107 | +```text |
| 108 | +Use the GitHub MCP server to search for recent issues in the microsoft/vscode repository |
| 109 | +``` |
| 110 | + |
| 111 | +If you have the GitHub MCP server running, the AI will execute the appropriate |
| 112 | +GitHub API calls and return formatted results. |
| 113 | + |
| 114 | +### Managing servers through conversation |
| 115 | + |
| 116 | +The ToolHive desktop app automatically starts a dedicated MCP server |
| 117 | +(`toolhive mcp`) that orchestrates ToolHive operations through natural language |
| 118 | +commands. This approach provides several key benefits: |
| 119 | + |
| 120 | +- **Unified interface**: Manage your MCP infrastructure using the same |
| 121 | + conversational AI interface you use for testing |
| 122 | +- **Contextual operations**: The AI understands your current server state and |
| 123 | + can make intelligent decisions about which servers to start, stop, or |
| 124 | + troubleshoot |
| 125 | +- **Reduced complexity**: No need to switch between the chat interface and |
| 126 | + traditional UI controls—everything can be done through conversation |
| 127 | +- **Audit trail**: All management operations are logged in the same transparent |
| 128 | + way as tool executions, providing clear visibility into what actions were |
| 129 | + taken |
| 130 | + |
| 131 | +Take advantage of these integrated ToolHive management tools: |
| 132 | + |
| 133 | +```text |
| 134 | +Start the fetch MCP server for me |
| 135 | +``` |
| 136 | + |
| 137 | +```text |
| 138 | +Stop all unhealthy MCP servers |
| 139 | +``` |
| 140 | + |
| 141 | +```text |
| 142 | +Show me the logs for the meta-mcp server |
| 143 | +``` |
| 144 | + |
| 145 | +### Validating tool responses |
| 146 | + |
| 147 | +The playground shows detailed information about each tool execution: |
| 148 | + |
| 149 | +- **Tool name and description**: What tool was called and its purpose |
| 150 | +- **Input parameters**: The exact parameters passed to the tool |
| 151 | +- **Execution status**: Whether the tool succeeded or failed |
| 152 | +- **Response data**: The complete response from the tool |
| 153 | +- **Timing information**: How long the tool took to execute |
| 154 | + |
| 155 | +This visibility helps you understand exactly how your MCP servers are behaving |
| 156 | +and identify any issues with tool implementation or configuration. |
| 157 | + |
| 158 | +## Recommended practices |
| 159 | + |
| 160 | +### API key security |
| 161 | + |
| 162 | +- Use dedicated API keys for testing that have appropriate rate limits |
| 163 | +- Regularly rotate API keys used in development environments |
| 164 | +- Consider using API keys with restricted permissions for testing purposes |
| 165 | + |
| 166 | +### Server management |
| 167 | + |
| 168 | +- Start only the MCP servers you need for testing to improve performance |
| 169 | +- Use the playground to validate new server configurations before connecting |
| 170 | + them to production AI clients |
| 171 | +- Test different combinations of tools to understand how they work together |
| 172 | + |
| 173 | +### Testing workflow |
| 174 | + |
| 175 | +1. **Isolated testing**: Test individual MCP servers one at a time to validate |
| 176 | + their functionality |
| 177 | +2. **Integration testing**: Enable multiple servers to test how they work |
| 178 | + together |
| 179 | +3. **Performance validation**: Monitor tool execution times and responses under |
| 180 | + different loads |
| 181 | +4. **Error handling**: Intentionally trigger error conditions to validate proper |
| 182 | + error handling |
| 183 | + |
| 184 | +## Troubleshooting |
| 185 | + |
| 186 | +### Common issues |
| 187 | + |
| 188 | +<details> |
| 189 | +<summary>API key not working</summary> |
| 190 | + |
| 191 | +If your API key isn't working: |
| 192 | + |
| 193 | +1. Verify the API key is correct and has proper permissions |
| 194 | +2. Check that you have sufficient quota or credits with the provider |
| 195 | +3. Ensure the API key hasn't expired or been revoked |
| 196 | +4. Try creating a new API key from the provider's dashboard |
| 197 | + |
| 198 | +</details> |
| 199 | + |
| 200 | +<details> |
| 201 | +<summary>MCP tools not appearing</summary> |
| 202 | + |
| 203 | +If your MCP server tools aren't showing up: |
| 204 | + |
| 205 | +1. Verify the MCP server is running (check the **MCP Servers** page) |
| 206 | +2. Click the tools icon and ensure the server's tools are enabled |
| 207 | +3. Try restarting the MCP server if it shows as unhealthy |
| 208 | +4. Check the server logs for any error messages |
| 209 | + |
| 210 | +</details> |
| 211 | + |
| 212 | +<details> |
| 213 | +<summary>Tool execution failing</summary> |
| 214 | + |
| 215 | +If tools are failing to execute: |
| 216 | + |
| 217 | +1. Check the tool's parameter requirements in the audit log |
| 218 | +2. Verify any required secrets or environment variables are configured |
| 219 | +3. Ensure the MCP server has necessary permissions (network access, file system |
| 220 | + access) |
| 221 | +4. Review the server logs for detailed error information |
| 222 | + |
| 223 | +</details> |
| 224 | + |
| 225 | +## Next steps |
| 226 | + |
| 227 | +- Learn about [client configuration](./client-configuration.mdx) to connect |
| 228 | + ToolHive to external AI applications |
| 229 | +- Set up [secrets management](./secrets-management.md) for secure handling of |
| 230 | + API keys and tokens |
| 231 | +- Explore [network isolation](./network-isolation.mdx) for enhanced security |
| 232 | + when testing untrusted MCP servers |
| 233 | +- Browse the [registry](./registry.mdx) to discover new MCP servers to test in |
| 234 | + the playground |
| 235 | + |
| 236 | +## Related information |
| 237 | + |
| 238 | +- [Run MCP servers](./run-mcp-servers.mdx) |
| 239 | +- [Install ToolHive](./install.mdx) |
0 commit comments