Common MCP Clients
Model Context Protocol (MCP) is an open protocol that enables clients such as code editors, browsers, and other developer tools to send structured, real-time context to AI models and copilots. MCP improves the usefulness and accuracy of AI assistants by supplying them with relevant information such as file trees, userCan use the system on a limited basis based on the rights assigned by the admin. selections, metadata, and more.
Connection Methods
Boost.spaceA platform that centralizes and synchronizes company data from internal and external sources, offering a suite of modules and addons for project management, CRM, data visualization, and more. Has many features to optimize your workflow! MCP supports two connectionUnique, active service acces point to a network. There are different types of connections (API key, Oauth…). types:
- ✅ OAuth (Recommended) – Simple, secure, and handled directly inside the AI client.
- ⚙️ URL + MCP TokenA secure code used to authenticate and authorize access to API endpoints, enabling users to connect with third-party applications. – Manual configuration for advanced usersCan use the system on a limited basis based on the rights assigned by the admin.. Availability depends on the specific AI tool.
We strongly recommend using OAuth whenever supported.
1. OpenAI
Step-by-step guide: Connect Boost.space to OpenAI via OAuth
Step 1: Copy your MCP system URL
1. Open Boost.space.
2. Navigate to Profile Settings → API / MCP Token → MCP Token.
3. Copy your dedicated system URL.
This is the only information required from Boost.space.
Step 2: Enable Developer Mode
1. Open the OpenAI platform.
2. Go to Apps → Advanced Settings.
3. Enable Developer Mode.
4. Click Create App.
Step 3: Create the MCP connection
1. Paste your Boost.space systemA platform that centralizes and synchronizes company data from internal and external sources, offering a suite of modules and addons for project management, CRM, data visualization, and more. Has many features to optimize your workflow! URL.
2. Give the app a clear name.
3. Review the permission notice.
4. Click Create.
Step 4: Activate in Chat
1. Open Chat window.
2. Click the plus icon.
3. Select your newly created MCP connection.
The model can now securely interact with your Boost.space system within the defined scope.
2. Anthropic Claude
Step-by-step guide: Connect Boost.space to Claude via OAuth
Step 1: Copy your MCP system URL
1. Open Boost.space.
2. Go to Profile Settings → API /MCP Token → MCP Token.
3. Copy your dedicated system URL.
Step 2: Add Custom Connector
1. Open Claude.
2. Click your avatar (bottom left).
3. Go to Settings → Connectors. Click Add Custom Connector.
4. Name the connection. Paste the system URL.
5. Click Add. Click Connect to authorize.
Step 3: Activate in Chat
1. Open a new chat.
2. Click the plus icon.
3. Select Connectors. Choose your MCP connection.
Claude can now retrieve Boost.space data or triggerThe initial event that starts a scenario. It can be an action, a scheduled time, or a custom event, and is essential to define at the beginning of each scenario. permitted scenariosA specific connection between applications in which data can be transferred. Two types of scenarios: active/inactive..
3. Lovable
Lovable supports both connection methods.
Supported connection types:
- ✅ OAuth (Recommended)
- ⚙️ URL + MCP Token
OAuth is recommended for most users.
Step-by-step guide: Connect Boost.space to Lovable via OAuth
Step 1: Copy your MCP system URL
1. Open Boost.space.
2. Navigate to Profile Settings → API / MCP Token → MCP Token.
3. Copy your system URL.
Step 2: Add MCP Connector in Lovable
1. Log in to Lovable.
2. Click your avatar → Settings. Open Connectors.
3. Scroll to Personal Connectors → Custom.
4. Enter a name. Paste your Boost.space system URL.
5. Click Add.
Once confirmed, the connector is ready.
Step 3: Activate in Chat
1. Open Chat.
2. Click the plus icon.
3. Choose Integrations → MCP Servers.
4. Select your MCP connection.
Lovable can now work directly with your Boost.space data.
Final Recommendation
Whenever possible, use OAuth.
It is:
-
Easier to configure
-
More secure
-
Fully integrated into modern AI interfaces
The URL + MCP token method remains available for advanced users or specific desktop environments, but it should be considered a secondary option.
If your AI client supports OAuth, that should always be your first choice.
If you still need to clarify anything, please reach out to us at [email protected].