MCP Server

Model Context Protocol (MCP) Server

Connect AI agents to your personalized context, skills, and workflows. Stop building orchestration. Start shipping agents.

What is MCP?

The Model Context Protocol (MCP) is an open standard for connecting AI agents to external data sources and tools. ContextFile.ai implements an MCP server that exposes your personalized context, skills, tasks, and use cases to any MCP-compatible AI client.

AI Agent Integration

Connect Claude Desktop, custom agents, or any MCP client to your personalized AI workflows.

Automated Workflows

Execute tasks, skills, and use cases programmatically through standardized MCP tools.

Personalized Context

Give AI agents access to your context files for personalized, contextually-aware responses.

Getting Started

1Get an API Key

Navigate to API Keys in your dashboard and create a new API key. Make sure to save it securely - you'll only see it once.

Go to API Keys

2MCP Server Endpoint

The MCP server is available at the following endpoint:

https://contextfile.ai/api/mcp

3Configure Your Client

Add the MCP server to your AI client configuration (see examples below).

Manual Configuration

If you prefer to configure manually, use these examples for different clients.

Add to your Claude Desktop configuration file:

{
  "mcpServers": {
    "contextfile": {
      "url": "https://contextfile.ai/api/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_API_KEY"
      }
    }
  }
}

Config file locations:
macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%\Claude\claude_desktop_config.json

Available Tools

Tools are executable actions that AI agents can invoke. Each tool requires specific scopes for authorization.

get_contextcontexts:read

Get the user's context file content

Params: context_id (optional)

list_contextscontexts:read

List all context files for the user

Params: limit, context_type

search_contextscontexts:read

Search across user's context files

Params: query (required), limit

get_skillskills:read

Load a specific skill by ID or name

Params: skill_id, skill_name

list_skillsskills:read

List all available skills for the user

Params: limit, category, active_only

list_taskstasks:read

List all available tasks for the user

Params: limit, category, active_only

run_tasktasks:execute

Execute a task with input

Params: task_id (required), input (required), variables, context_id

execute_tasktasks:execute

Execute a task workflow with variables

Params: task_id (required), variables, context_id

run_use_caseuse_cases:execute

Execute a complete use case workflow

Params: use_case_id (required), input (required), variables

Available Resources

Resources provide read-only access to data. AI agents can read resources to get context, skills, tasks, and more.

context://primary

The user's primary context file

context://all

List of all user context files

context://{id}

Specific context file by ID

skills://active

List of active user skills

skill://{id}

Specific skill by ID

tasks://active

List of active user tasks

task://{id}

Specific task by ID

use-cases://active

List of active use cases

use-case://{id}

Specific use case by ID

profile://info

User profile information

Available Prompts

Prompts are pre-configured templates that generate contextual messages. Dynamic prompts are automatically created from your active use cases.

context-summary

Get a summary of your context file

Args: context_id (optional)

skill-guide

Get guidance on using a specific skill

Args: skill_name (required)

task-instructions

Get detailed instructions for a task

Args: task_id (required), include_context

introduction

Generate a professional introduction

Args: audience, length

expertise-areas

List key expertise areas from context

Args: none

use-case-assemble

Assemble a complete prompt from a use case

Args: use_case_id (required), variables

use-case-guide

Get guidance on a use case workflow

Args: use_case_id (required)

use-case-list

List available use cases

Args: category, active_only

usecase:{id}Dynamic

Dynamic prompt - execute any active use case

Args: input (required), variables, context_id

Dynamic Prompts

Each active use case in your account automatically becomes a dynamic prompt with the format usecase:{id}. This lets you execute any use case as a prompt directly.

Example Usage

Execute a tool to get your primary context:

// Request
{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "get_context",
    "arguments": {}
  }
}

// Response
{
  "jsonrpc": "2.0",
  "id": 1,
  "result": {
    "content": [{
      "type": "text",
      "text": "{ \"id\": \"..\", \"name\": \"My Context\", ... }"
    }],
    "isError": false
  }
}

Execute a use case workflow:

// Request
{
  "jsonrpc": "2.0",
  "id": 2,
  "method": "tools/call",
  "params": {
    "name": "run_use_case",
    "arguments": {
      "use_case_id": "write-blog-post",
      "input": "How to use MCP with ContextFile.ai",
      "variables": {
        "tone": "professional",
        "length": "medium"
      }
    }
  }
}

// Response includes the generated content
{
  "jsonrpc": "2.0",
  "id": 2,
  "result": {
    "content": [{
      "type": "text",
      "text": "{ \"use_case\": {...}, \"status\": \"completed\", ... }"
    }],
    "isError": false
  }
}

Authentication & Scopes

All MCP requests require authentication via API key. Each API key can have specific scopes that control what operations it can perform.

Available Scopes

contexts:read - Read context files
skills:read - Read skills
tasks:read - Read tasks
tasks:execute - Execute tasks
use_cases:execute - Execute use cases
* - Full access (default for session auth)

Authentication Methods

API Key: Authorization: Bearer cf_xxx
OAuth Token: Authorization: Bearer cf_at_xxx
Session: Cookie-based (browser requests)

Tips for Building with MCP

Use get_context to provide AI agents with your personalized context automatically
Combine skills and tasks to create powerful automated workflows
Use run_use_case for complex multi-step operations
Dynamic prompts (usecase:{id}) let you execute any use case as a prompt
Resources provide read-only access - great for displaying data
Scopes control what operations each API key can perform

Ready to connect AI agents to your context?

Get API Key Learn About Tasks Create Use Cases

Docs Dashboard

MCP Server

Model Context Protocol (MCP) Server

Connect AI agents to your personalized context, skills, and workflows. Stop building orchestration. Start shipping agents.

What is MCP?

AI Agent Integration

Connect Claude Desktop, custom agents, or any MCP client to your personalized AI workflows.

Automated Workflows

Execute tasks, skills, and use cases programmatically through standardized MCP tools.

Personalized Context

Give AI agents access to your context files for personalized, contextually-aware responses.

Getting Started

1Get an API Key

Navigate to API Keys in your dashboard and create a new API key. Make sure to save it securely - you'll only see it once.

Go to API Keys

2MCP Server Endpoint

The MCP server is available at the following endpoint:

https://contextfile.ai/api/mcp

3Configure Your Client

Add the MCP server to your AI client configuration (see examples below).

Manual Configuration

If you prefer to configure manually, use these examples for different clients.

Add to your Claude Desktop configuration file:

{
  "mcpServers": {
    "contextfile": {
      "url": "https://contextfile.ai/api/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_API_KEY"
      }
    }
  }
}

Config file locations:
macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%\Claude\claude_desktop_config.json

Available Tools

Tools are executable actions that AI agents can invoke. Each tool requires specific scopes for authorization.

get_contextcontexts:read

Get the user's context file content

Params: context_id (optional)

list_contextscontexts:read

List all context files for the user

Params: limit, context_type

search_contextscontexts:read

Search across user's context files

Params: query (required), limit

get_skillskills:read

Load a specific skill by ID or name

Params: skill_id, skill_name

list_skillsskills:read

List all available skills for the user

Params: limit, category, active_only

list_taskstasks:read

List all available tasks for the user

Params: limit, category, active_only

run_tasktasks:execute

Execute a task with input

Params: task_id (required), input (required), variables, context_id

execute_tasktasks:execute

Execute a task workflow with variables

Params: task_id (required), variables, context_id

run_use_caseuse_cases:execute

Execute a complete use case workflow

Params: use_case_id (required), input (required), variables

Available Resources

Resources provide read-only access to data. AI agents can read resources to get context, skills, tasks, and more.

context://primary

The user's primary context file

context://all

List of all user context files

context://{id}

Specific context file by ID

skills://active

List of active user skills

skill://{id}

Specific skill by ID

tasks://active

List of active user tasks

task://{id}

Specific task by ID

use-cases://active

List of active use cases

use-case://{id}

Specific use case by ID

profile://info

User profile information

Available Prompts

Prompts are pre-configured templates that generate contextual messages. Dynamic prompts are automatically created from your active use cases.

context-summary

Get a summary of your context file

Args: context_id (optional)

skill-guide

Get guidance on using a specific skill

Args: skill_name (required)

task-instructions

Get detailed instructions for a task

Args: task_id (required), include_context

introduction

Generate a professional introduction

Args: audience, length

expertise-areas

List key expertise areas from context

Args: none

use-case-assemble

Assemble a complete prompt from a use case

Args: use_case_id (required), variables

use-case-guide

Get guidance on a use case workflow

Args: use_case_id (required)

use-case-list

List available use cases

Args: category, active_only

usecase:{id}Dynamic

Dynamic prompt - execute any active use case

Args: input (required), variables, context_id

Dynamic Prompts

Each active use case in your account automatically becomes a dynamic prompt with the format usecase:{id}. This lets you execute any use case as a prompt directly.

Example Usage

Execute a tool to get your primary context:

// Request
{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "get_context",
    "arguments": {}
  }
}

// Response
{
  "jsonrpc": "2.0",
  "id": 1,
  "result": {
    "content": [{
      "type": "text",
      "text": "{ \"id\": \"..\", \"name\": \"My Context\", ... }"
    }],
    "isError": false
  }
}

Execute a use case workflow:

// Request
{
  "jsonrpc": "2.0",
  "id": 2,
  "method": "tools/call",
  "params": {
    "name": "run_use_case",
    "arguments": {
      "use_case_id": "write-blog-post",
      "input": "How to use MCP with ContextFile.ai",
      "variables": {
        "tone": "professional",
        "length": "medium"
      }
    }
  }
}

// Response includes the generated content
{
  "jsonrpc": "2.0",
  "id": 2,
  "result": {
    "content": [{
      "type": "text",
      "text": "{ \"use_case\": {...}, \"status\": \"completed\", ... }"
    }],
    "isError": false
  }
}

Authentication & Scopes

All MCP requests require authentication via API key. Each API key can have specific scopes that control what operations it can perform.

Available Scopes

contexts:read - Read context files
skills:read - Read skills
tasks:read - Read tasks
tasks:execute - Execute tasks
use_cases:execute - Execute use cases
* - Full access (default for session auth)

Authentication Methods

API Key: Authorization: Bearer cf_xxx
OAuth Token: Authorization: Bearer cf_at_xxx
Session: Cookie-based (browser requests)

Tips for Building with MCP

Use get_context to provide AI agents with your personalized context automatically
Combine skills and tasks to create powerful automated workflows
Use run_use_case for complex multi-step operations
Dynamic prompts (usecase:{id}) let you execute any use case as a prompt
Resources provide read-only access - great for displaying data
Scopes control what operations each API key can perform

Ready to connect AI agents to your context?

Get API Key Learn About Tasks Create Use Cases