# AIsuru MCP: which ones and how to activate them Here you'll learn how to correctly configure the parameters for each MCP server available on AIsuru. ## How to configure an MCP To configure an MCP in your Agent: 1. Go to your Agent's management panel; 2. From the sidebar, go to the MCP section; 3. Press on the MCP you want to activate; 4. Fill in the required fields; 5. Click "Create MCP" to save the configuration. Now let's see how to create an MCP. ## 0. Creating a custom MCP Use this process if the MCP you need isn't among the pre-configured ones. After clicking "Add a custom MCP", a modal window titled "Create New MCP" will open with several fields to fill in. #### Configuration parameters: * **Name (required):** type an identifying name (e.g., "Custom MCP Server", "CRM Integration"); * **Description (required):** type a detailed description of what the MCP does; * **Prompt (optional):** enter instructions on how the agent should use this MCP; * **Authentication token (optional):** paste the authentication token from your external service here; * **URL (required):** enter the identifying URL of your MCP server. #### Saving the custom MCP 1. Make sure all required fields (Name, Description, URL) are filled in; 2. Click the "Create MCP" button in the bottom right; 3. Wait for the confirmation message. To cancel: click "Cancel" or the X in the top right. ## 1. AIsuru MCP Server Enables direct integration with the AIsuru platform, letting your Agent communicate with other agents or platform features. This MCP is particularly useful for: * Creating multi-agent architectures; * Enabling communication between different agents; * Accessing advanced AIsuru platform features. #### Configuration parameters * **Backend URL**: ; * **Engine URL**: ; * **Tenant URL**: aisuru.com (SaaS) or your tenant domain ([PaaS](/en/paas/paas-benefits.md)). #### How to use the AIsuru MCP to have two Agents talk to each other If you want your Agent to consult a specific agent, write it in the [instructions](/en/generative-ai/ai-instructions.md) — for example: if you want to contact "Manuela" for HR questions, write in the prompt: *"If the user asks about HR, direct the question to Manuela (ID: b6fc6009-529f-42c1-91e6-5228f695fc81)"*. {% hint style="info" %} The ID can be retrieved in a few simple steps: * Select and open an agent's side panel; * Under Settings, press "Dev docs"; * Once inside dev docs you'll see the agent's ID and a button to copy it to the clipboard. {% endhint %} ## 2. Free Image Generator The Free Image Generator lets your Agent generate images based on text descriptions using the Pollinations.ai service. This MCP is useful for: * Creating custom images during conversations; * Generating visual content based on user requests; * Enriching responses with graphic elements. #### Configuration parameters This MCP doesn't require mandatory parameters. The basic configuration is enough to start generating images. #### Custom parameters * **timeout**: Maximum time (in seconds) to generate the image before aborting the request; * **retries**: Number of automatic retries in case of errors during generation; * **defaultWidth**: Default image width in pixels if not specified by the user; * **defaultHeight**: Default image height in pixels if not specified by the user; * **maxSize**: Maximum allowed size (in pixels) for image width and height. ```` ```json { "timeout": 60, "retries": 3, "defaultWidth": 1024, "defaultHeight": 1024, "maxSize": 2048 } ``` ```` {% hint style="info" %} Check the Terms and Conditions on the Pollinations.ai website before using this MCP in production. {% endhint %} ## 3. MCP Persistence Gives your Agent a **personal database** to save and retrieve information persistently across conversations. This MCP is particularly useful for: * Storing user preferences and data; * Creating a personalized knowledge base; * Saving history and context across different sessions; * Managing archives of documents and structured information. #### Configuration parameters This MCP doesn't require mandatory parameters. The basic configuration is enough to start saving data. #### How to use MCP Persistence Add instructions in the prompt to tell the Agent **what to save and when to retrieve** data. **Practical examples:** ``` Save user preferences to the database whenever the user specifies them. Before responding, check the database for any existing information about the current user. ``` The Agent will automatically manage read/write operations on the database through the MCP. ## 4. MongoDB MCP Server The MongoDB MCP Server lets your Agent connect to a MongoDB database and perform read and write operations. This MCP is useful for: * Querying MongoDB databases; * Storing and retrieving data during conversations; * Managing structured data in document format. #### Parameters | Parameter | Description | Format | Example | | ------------------- | ------------------------------ | -------------------------------- | ------------------------------------------------------------------------------------------------------------- | | `Connection String` | Full MongoDB connection string | `mongodb://` or `mongodb+srv://` |

mongodb://localhost:27017/
mongodb+srv://user:/

| | `Database Name` | Name of the database to use | Alphanumeric string | `aisuru_test`, `production_db` | #### How to get the parameters 1. **After registering**, create a cluster (choose the M0 Free plan); 2. Click **"Connect"** on your cluster; 3. Choose **"Connect your application"**; 4. Select **Driver**: Node.js or Python; 5. Copy the **Connection String**: ``` mongodb+srv://:@cluster0.xxxxx.mongodb.net/ ``` 6. Replace `` with your database user's password; 7. **Database Name**: choose one (e.g., `test_db`, `aisuru`). #### Custom parameters * **timeout**: Maximum time to complete an operation (ms); * **maxPoolSize**: Maximum number of simultaneous connections in the pool; * **minPoolSize**: Minimum number of always-active connections; * **retryWrites**: Automatically retry failed writes; * **readPreference**: Which node to read data from (primary/secondary/nearest); * **maxIdleTimeMS**: Maximum idle time before closing a connection (ms); * **serverSelectionTimeoutMS**: Maximum time to find an available server (ms). ````json ```json { "timeout": 30000, "maxPoolSize": 10, "minPoolSize": 2, "retryWrites": true, "readPreference": "primary", "maxIdleTimeMS": 60000, "serverSelectionTimeoutMS": 30000 } ``` ```` ## 5. AIsuru Data Analysis AIsuru Data Analysis is a JavaScript MCP that supports data analysis with an isolated execution environment. It lets your Agent run JavaScript/TypeScript code for complex analysis, processing, and business intelligence. This MCP is useful for: * Analyzing complex datasets; * Performing calculations and aggregations; * Processing data in real time during conversations. #### Configuration parameters This MCP doesn't require mandatory parameters in the basic configuration. The execution environment is provided automatically. #### Custom parameters * **timeout**: Maximum operation time (seconds); * **allowedFormats**: Accepted file formats. ````json ```json { "timeout": 120, "allowedFormats": ["json", "csv", "xlsx"] } ``` ```` #### Important notes * Code runs in an isolated environment (isolated-vm); * Maximum timeout is limited to 128MB; * Access to external URLs is limited; * Code can handle both simple expressions and complex code with variables and refactoring; * Always check memory and timeout limits before using this MCP for intensive analysis. ## 6. PostgreSQL MCP Server The PostgreSQL MCP Server lets your Agent connect to PostgreSQL databases and perform read and write operations. This MCP is useful for: * Querying relational PostgreSQL databases; * Running complex SQL queries; * Managing structured data in relational tables. #### Required parameters | Parameter | Description | Format | Example | | ------------------- | --------------------------------- | ------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------- | | `Connection String` | Full PostgreSQL connection string | `postgresql://user:pass@host:port/db` |

postgresql://postgres:pass\@localhost:5432/mydb
postgresql://postgres::5432/postgres

| #### Optional parameters | Parameter | Description | Default | Possible values | | ------------------------- | -------------------------------------------- | ------------ | ------------------------------------------------------------------- | | `Access Mode` | Database access mode | `read-write` | `read-write`, `read-only` | | `Ssl Mode` | SSL mode for the connection | `prefer` | `disable`, `allow`, `prefer`, `require`, `verify-ca`, `verify-full` | | `Ssl Cert Path` | Path to the SSL client certificate | - | `/path/to/client-cert.pem` | | `Ssl Key Path` | Path to the SSL private key | - | `/path/to/client-key.pem` | | `Ssl Ca Path` | Path to the CA certificate | - | `/path/to/ca-cert.pem` | | `Ssl Reject Unauthorized` | Reject connections with invalid certificates | `true` | `true`, `false` | #### How to get the parameters 1. **After registering**, create a new project; 2. Go to **Settings** (gear icon) → **Database**; 3. Copy the **Connection string** in URI format: ``` postgresql://postgres:[YOUR-PASSWORD]@db.xxxxx.supabase.co:5432/postgres ``` 4. **Ssl Mode**: use `require` (Supabase requires SSL); **Note on Supabase Connection Pooler**: * **Port 5432**: Direct connection (for transactions and complex operations); * **Port 6543**: Connection pooler (for applications with many simultaneous connections). #### Custom parameters * **poolSize**: Total number of connections in the pool; * **connectionTimeout**: Maximum time to establish a connection (seconds); * **idleTimeout**: Maximum idle time for connections (seconds); * **queryTimeout**: Maximum time to execute a query (seconds); * **statementTimeout**: Maximum time for an SQL statement (ms); * **applicationName**: Identifying name of the application. ````json ```json { "poolSize": 20, "connectionTimeout": 30, "idleTimeout": 60, "queryTimeout": 60, "statementTimeout": 60000, "applicationName": "aisuru_mcp" } ``` ```` ## 7. n8n Workflow Server The n8n Workflow Server lets your Agent access the real tools of n8n workflows through supergateway. n8n is an automation platform that lets you create complex workflows by connecting different services. This MCP is useful for: * Integrating existing n8n workflows; * Automating complex processes; * Connecting your Agent to hundreds of services supported by n8n. #### Required parameters | Parameter | Description | Format | Example | | -------------- | ---------------------------------- | ------------------- | ----------------------------------------------------------------------------------------------------------------------- | | `Workflow Url` | URL of the n8n webhook or endpoint | `https://` full URL |


| #### Optional parameters | Parameter | Description | Default | Example | | ------------------- | ------------------------------------------------- | ------- | ------------------------------------- | | `Workflow Name` | Identifying name of the workflow | - | `process_orders`, `sync_customers` | | `Supergateway Args` | Additional arguments for supergateway NPX | - | Specific parameters for NPX execution | | `Auth Token` | Authentication token (if the webhook requires it) | - | `Bearer token123...` | #### How to get the parameters **Option A: n8n Cloud** 1. **After registering** and logging in to n8n Cloud 2. **Create a workflow** with a **Webhook** node 3. In the Webhook node: * Click **"Webhook URLs"** * Copy the **production** or **test** URL #### **Custom parameters:** * **timeout**: Maximum time for a request (seconds); * **retries**: Number of retries in case of failure; * **retryDelay**: Pause between retries (ms); * **headers**: Custom HTTP headers to send. ````json ```json { "timeout": 60, "retries": 3, "retryDelay": 1000, "headers": { "X-Custom-Header": "value" } } ``` ```` ## 8. Fetch URL Fetch URL is a tool for retrieving information from a link and converting it to HTML or markdown format. This MCP is useful for: * Extracting content from web pages; * Retrieving data from public APIs; * Converting web content into formats the Agent can process. #### Configuration parameters This MCP doesn't require mandatory parameters. The basic configuration lets you retrieve content from any public URL. #### Important notes For more information on how this MCP works, see the documentation on GitHub: `https://github.com/modelcontextprotocol/servers/tree/main/src/fetch`. ## 9. Microsoft SQL Server MCP The Microsoft SQL Server MCP lets your Agent connect to Microsoft SQL Server and Azure SQL databases. Supports SQL queries, schema management, metrics, and enterprise security. This MCP is useful for: * Querying SQL Server databases; * Running complex queries on enterprise databases; * Managing data in Microsoft Azure environments. #### Required parameters | Parameter | Description | Format | Example | | ------------- | --------------------------- | -------------------------------------------------------- | --------------------------------- | | `DB_HOST` | SQL server address | `servername.database.windows.net` (Azure) or `localhost` | `aisuru-sql.database.windows.net` | | `DB_DATABASE` | Database name | Alphanumeric string | `aisuru_db`, `production` | | `DB_USER` | Username for authentication | `admin` or `admin@servername` | `sqladmin` | | `DB_PASSWORD` | User password | Complex string | `MySecureP@ssw0rd123!` | #### **Optional parameters** | Parameter | Description | Default | Possible values | | ----------------------------- | ------------------------------- | ------- | --------------- | | `DB_PORT` | SQL server port | `1433` | Any TCP port | | `DB_ENCRYPT` | Enable TLS/SSL encryption | `true` | `true`, `false` | | `DB_TRUST_SERVER_CERTIFICATE` | Accept self-signed certificates | `false` | `true`, `false` | | `CONNECTION_TIMEOUT` | Connection timeout in seconds | `15` | 5–120 | | `REQUEST_TIMEOUT` | Query timeout in seconds | `15` | 5–300 | | `POOL_SIZE` | Connection pool size | `10` | 1–100 | #### How to get the parameters **Option A: Azure SQL Database (Cloud)** 1. **After creating** an Azure SQL Database; 2. Go to **Azure Portal** → **SQL databases** → select your database; 3. In the side menu, click **Connection strings**; 4. You'll find a string like: ``` Server=tcp:myserver.database.windows.net,1433; Initial Catalog=mydatabase; User ID=sqladmin; Password={your_password}; Encrypt=True; ``` 5. **Extract the parameters**: * `DB_HOST`: `myserver.database.windows.net` * `DB_DATABASE`: `mydatabase` * `DB_USER`: `sqladmin` * `DB_PASSWORD`: the password you set * `DB_PORT`: `1433` (or leave blank to use default) * `DB_ENCRYPT`: `true` * `DB_TRUST_SERVER_CERTIFICATE`: `false` #### Custom parameters * **connectionTimeout**: Maximum time to connect (seconds); * **requestTimeout**: Maximum time for a request (seconds); * **poolSize**: Number of connections in the pool; * **enableArithAbort**: Enable arithmetic error interruption; * **trustServerCertificate**: Accept self-signed SSL certificates. ````json ```json { "connectionTimeout": 15, "requestTimeout": 30, "poolSize": 10, "enableArithAbort": true, "options": { "trustServerCertificate": false, "enableArithAbort": true } } ``` ```` ## 10. Monday.com MCP Server The Monday.com MCP Server lets your Agent interact with the Monday.com platform to manage projects, tasks, and team collaboration. This MCP is useful for: * Creating and updating tasks in Monday.com; * Retrieving information about projects; * Automating work management. #### Parameters | Parameter | Description | Format | Example | | ---------------- | ------------------------------------ | ------------------------ | ---------------------------------- | | `Workspace Name` | Monday.com workspace name | String | `Main Workspace`, `Marketing Team` | | `Api Key` | Authentication API token | Long alphanumeric string | `eyJhbGciOiJIUzI1NiJ9...` | | `Token Param` | Parameter name for passing the token | String | `Authorization`, `api_token` | #### How to get the parameters 1. **After registering** and logging in to Monday.com 2. **Workspace Name**: * Look in the left sidebar; * You'll see the current workspace name (e.g., "Main Workspace", "Marketing Team"); * Use that exact name; 3. **Api Key** (API Token): * Click your **profile photo** (bottom left); * Select **Admin** (if you're an admin) or **Developers**; * In the side menu, go to **Developers** → **API Token**; * Or go directly to: ; * In the **"API Token"** section: * Click **"Generate"** to create a new token; * Or **"Show"** to see an existing one; * Copy the token (often starts with `eyJ...`); {% hint style="warning" %} **IMPORTANT**: Save the token securely! {% endhint %} 1. **Token Param**: * This specifies how to pass the token in API requests; * For Monday.com, use: **`Authorization`**; * The standard format is: `Authorization: Bearer {token}`. #### Complete configuration example ``` Workspace Name: Main Workspace Api Key: eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJzdWI... Token Param: Authorization ``` #### API Token security * The API token has **the same permissions** as the user who generated it; * It can access all of that user's boards and workspaces; * Never share the token publicly; * Rotate the token periodically for security. #### Additional information **Finding Workspace ID** (for reference): * URL: `https://mycompany.monday.com/workspaces/12345678` * The ID is `12345678` **Finding Board ID** (for reference): * URL: `https://mycompany.monday.com/boards/123456789` * The ID is `123456789` #### Custom parameters * **apiVersion**: API version to use; * **timeout**: Maximum request time (seconds); * **retries**: Retries in case of error; * **maxQueries**: GraphQL query complexity limit; * **resetInterval**: Complexity limit reset interval (seconds). ````json ```json { "apiVersion": "2024-01", "timeout": 60, "retries": 3, "complexity": { "maxQueries": 100, "resetInterval": 60 } } ``` ```` ## 11. Salesforce MCP Server The Salesforce MCP Server lets your Agent interact with Salesforce via NPX. It provides access to Salesforce data and features using OAuth 2.0 Client Credentials. This MCP is useful for: * Retrieving information from Salesforce CRM; * Updating records and opportunities; * Automating sales processes. #### **Parameters** | Parameter | Description | Format | Example | | ---------------------------- | ---------------------------------------------- | ------------------------------------- | ----------------------------------- | | `Salesforce Client Id` | OAuth 2.0 client identifier for authentication | `client_id_string` | `3MVG9xyz123abc...` | | `Salesforce Client Secret` | OAuth 2.0 client secret | `client_secret_string` | `a1b2c3d4e5f6...` | | `Salesforce Instance Url` | URL of your Salesforce instance | `https://yourinstance.salesforce.com` | `https://company.my.salesforce.com` | | `Salesforce Connection Type` | The type of Salesforce connection to use | `oauth2` | `oauth2` | #### **Optional parameters** | Parameter | Description | Default | Example | | ------------------- | -------------------------------------------------------------------------- | ------- | ------------------------------------------------------- | | `Custom parameters` | Additional parameters in JSON format for timeout, retry, API version, etc. | `-` | `{"timeout": 30, "retries": 3, "api_version": "v58.0"}` | #### Custom parameters * **apiVersion**: Salesforce API version; * **timeout**: Maximum request time (seconds); * **retries**: Retries in case of error; * **loginUrl**: Authentication URL (different for sandbox); * **maxRecords**: Maximum number of records per query. ```` ```json { "apiVersion": "v60.0", "timeout": 60, "retries": 3, "loginUrl": "https://login.salesforce.com", "maxRecords": 2000 } ``` ```` ## 12. Wikipedia The Wikipedia MCP lets your Agent directly access Wikipedia content to retrieve encyclopedic information. This MCP is useful for: * Enriching responses with encyclopedic information; * Providing context on people, places, or events; * Accessing verified, structured knowledge. #### Configuration parameters This MCP doesn't require mandatory parameters. Wikipedia access is public and requires no authentication. #### Custom parameters * **timeout**: Maximum request time (seconds); * **language**: Wikipedia language code (it, en, es, fr...); * **maxResults**: Maximum number of search results; * **autoSuggest**: Enable automatic search suggestions. ```` ```json { "timeout": 30, "language": "en", "maxResults": 10, "autoSuggest": true } ``` ```` #### Important notes The Agent can use Wikipedia to retrieve updated information during conversations. Content is subject to Wikipedia's licenses and terms of use. ## 13. Zapier MCP Server The Zapier MCP Server lets your Agent access Zapier's real tools via supergateway. Zapier lets you automate workflows by connecting thousands of different applications. This MCP is useful for: * Integrating Zapier into your Agent; * Automating processes across different applications; * Connecting services that don't have dedicated MCPs. {% hint style="warning" %} In the configuration panel of your MCP on [Zapier MCP](https://mcp.zapier.com/), remember to:
* Enter "other" in the Client section; * In the "connect" section, set the transport to "SSE" {% endhint %} #### Required parameters | Parameter | Description | Format | Example | | -------------------- | ----------------------- | ------------------- | ---------------------------------------------------------------------------------------------------------------------------------- | | `Zapier Webhook Url` | Full Zapier webhook URL | `https://` full URL |



| #### **Optional parameters** | Parameter | Description | Default | Example | | ------------------- | ------------------------------------------------- | ------- | ------------------------------------- | | `Automation Name` | Identifying name for the Zapier automation | `-` | `process_orders`, `sync_customers` | | `Supergateway Args` | Additional arguments for supergateway NPX | `-` | Specific parameters for NPX execution | | `Auth Token` | Authentication token (if the webhook requires it) | `-` | `Bearer token123...` | #### Custom parameters * **timeout**: Maximum time for a request (seconds); * **retries**: Number of retries in case of failure; * **retryDelay**: Pause between retries (ms); * **headers**: Custom HTTP headers to send. ```` ```json { "timeout": 60, "retries": 3, "retryDelay": 1000, "headers": { "X-Custom-Header": "value" } } ``` ```` ### Best practices When configuring MCPs for your Agent, keep these tips in mind: * **Security**: never share API keys, passwords, or authentication tokens. Always use secure connections (HTTPS, SSL/TLS) when available; * **Testing**: always test the MCP configuration in a test environment before using it in production; * **Limits**: check the rate limits (requests per minute/hour) of external services to avoid getting blocked; * **Errors**: implement appropriate error handling in the Agent's instructions to deal with potential connection issues; * **Documentation**: always consult the official documentation of external services to understand available features and configuration requirements; * **Custom parameters**: use custom parameters in JSON format for advanced configurations, but always verify they're supported by the specific MCP. ### Limitations MCPs have some limitations to keep in mind: * **Timeout**: each request to an MCP has a maximum timeout, after which it's terminated; * **Rate limiting**: external services may limit the number of requests in a given period; * **Data security**: data passed through MCPs may be visible to external services. Don't use MCPs for sensitive data without checking the service's privacy policy; * **Availability**: MCPs depend on the availability of external services. If a service is offline, the MCP won't work; * **API versions**: some services may deprecate API versions. Keep your MCP configurations up to date. ### Troubleshooting common issues #### The MCP won't connect Check that: * All required parameters are filled in correctly; * Credentials (API key, password, token) are valid and not expired; * URLs are correct and reachable; * The external service is online and working. #### The Agent isn't using the MCP Make sure that: * The MCP was saved correctly after configuration; * The Agent's instructions include guidance on when and how to use the MCP; * The user's request is actually something the configured MCP can handle. #### Authentication errors Check that: * Credentials were copied correctly without extra spaces; * API keys haven't expired or been revoked; * The permissions associated with the credentials include the required operations. #### Timeout or slowness Consider: * Increasing the `timeout` value in the custom parameters; * Checking your network connection; * Optimizing queries or requests to reduce processing time. --- # Agent Instructions: Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter: ``` GET https://docs.aisuru.com/en/advanced-features/integrations/mcp/aisuru-mcp-which-ones-and-how-to-activate.md?ask= ``` The question should be specific, self-contained, and written in natural language. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.