๐ BolideAI MCP
BolideAI MCP is a comprehensive ModelContextProtocol (MCP) server that offers tools for marketing automation, content generation, research, and project management. It integrates with various AI services to streamline workflows for developers and marketers.
โจ Features
- ๐ Project Scaffolding: Create marketing project directories and structures.
- ๐ฑ Marketing Automation: Capture screenshots and videos using the companion app.
- ๐ค AI-Powered Content Generation: Generate social media posts using Gemini AI.
- ๐ Research Tools: Conduct comprehensive research using Perplexity AI and OpenAI.
- ๐ ๏ธ Diagnostic Tools: Perform system validation and troubleshooting.
๐ Quick Start
Quick Start (NPM Package)
For quick testing without local builds:
Installation
npm install -g @bolide-ai/mcp
Or use with npx (recommended):
npx --package=@bolide-ai/mcp bolide-ai-mcp
Configuration
Configure your MCP client with the following:
{
"mcpServers": {
"BolideAI": {
"command": "npx --package=@bolide-ai/mcp bolide-ai-mcp",
"env": {
"BOLIDEAI_API_KEY": "your-api-key"
}
}
}
}
Development Setup (Local Build)
For development or local builds, follow these detailed steps:
1. Install Prerequisites
Install Node.js 22+
# Download and install from https://nodejs.org/en/download
Clone the repositories
git clone https://github.com/Bolide-AI/mcp
2. Build the MCP Server
# In the directory of the cloned repository
npm install && npm run build
4. Configure MCP in Cursor
- In the Cursor menu, select Cursor โ Settings โ Cursor Settings.
- In the opened window, select Tools & Integrations.
- Click New MCP Server.
- Insert the MCP server configuration, replacing:
<PATH TO MCP DIRECTORY>with the path to the MCP directory.<BOLIDEAI_API_TOKEN>with your BolideAI key (can be generated at here).
{
"mcpServers": {
"BolideAI-dev": {
"type": "stdio",
"command": "node",
"args": [
"--inspect=9999",
"<PATH TO MCP DIRECTORY>/build/index.js"
],
"env": {
"BOLIDEAI_MCP_DEBUG": "true",
"BOLIDEAI_API_TOKEN": "<your-api-key-here>"
}
}
}
}
- Make sure the workspace is open in Cursor.
- Launch the application in the simulator.
๐ป Usage Examples
Basic Usage
// Create a project
scaffold_bolide_ai_project()
// Perform research
use_openai_deep_research({
query: "AI trends in marketing automation 2024"
})
// Analyze screencast content
analyze_screencasts({
screencastNames: ["demo.mov"],
force: false
})
๐ง Technical Details
Environment Variables
BOLIDEAI_API_TOKEN: Required for research tools.BOLIDEAI_API_URL: Optional, defaults to https://bolide.ai/api.
Tool Configuration
BOLIDEAI_MCP_DEBUG=true: Enable diagnostic tools and detailed logging.- Tool Groups: Enable specific tool categories (see Tool Options).
- Individual Tools: Enable specific tools only (see Tool Options).
๐ Documentation
Available Tools
BolideAI MCP provides 10 tools across 5 categories:
๐ Project Scaffolding
scaffold_bolide_ai_project: Create the bolide.ai project directory structure.
๐ฑ Utility Tools
check_companion_app_status: Check the companion app running status.launch_companion_app: Launch the Companion App for marketing capture.stop_companion_app: Stop running companion app instances.install_brew_and_ffmpeg: Install the Homebrew package manager and FFmpeg.
๐ค Content Generation
analyze_screencasts: Analyze screencast content using Gemini AI.generate_gif: Convert screencast segments to GIFs.enhance_audio: Extract and enhance audio from screencasts using ElevenLabs.
๐ Research Tools
use_perplexity: Conduct research using Perplexity AI.use_openai_deep_research: Perform deep research using OpenAI o4 - mini - deep - research.
๐ ๏ธ Diagnostic Tools
diagnostic: Perform system environment validation (debug mode only).
Research Tools
Perplexity AI Research
Perform quick research and information gathering:
use_perplexity({
query: "Latest trends in AI marketing automation",
search_mode: "web" // or "academic"
})
OpenAI Deep Research
Conduct comprehensive research with query enrichment:
use_openai_deep_research({
query: "Economic impact of renewable energy adoption"
})
The deep research tool:
- Enriches your query using GPT - 4.1 with detailed research instructions.
- Researches using o4 - mini - deep - research with web search and code interpreter.
- Returns both enriched instructions and comprehensive findings.
Common Workflows
Complete Marketing Content Creation
- Set up project structure:
scaffold_bolide_ai_project() - Check app status and capture content:
check_companion_app_status() launch_companion_app() - Enhance audio quality:
enhance_audio({ screencastNames: ["demo.mov"] })
Research Workflow
- Quick research:
use_perplexity({ query: "Your research question", search_mode: "web" }) - Deep research:
use_openai_deep_research({ query: "Complex research topic requiring comprehensive analysis" })
Configuration Options
Selective Tool Registration
Enable only the tools you need to optimize performance:
{
"env": {
"BOLIDEAI_MCP_GROUP_RESEARCH": "true",
"BOLIDEAI_MCP_GROUP_CONTENT_GENERATORS": "true",
"BOLIDEAI_API_TOKEN": "your-api-key"
}
}
Available Tool Groups
BOLIDEAI_MCP_GROUP_LAUNCH: Launch and utility tools.BOLIDEAI_MCP_GROUP_SCAFFOLDING: Project scaffolding tools.BOLIDEAI_MCP_GROUP_CONTENT_GENERATORS: Content generation tools.BOLIDEAI_MCP_GROUP_RESEARCH: Research and information gathering tools.BOLIDEAI_MCP_GROUP_DIAGNOSTICS: Diagnostic tools.
System Requirements
Dependencies
- Node.js 22+
- ffmpeg (required for GIF generation tools)
- Companion App
Troubleshooting
Enable Debug Mode
{
"env": {
"BOLIDEAI_MCP_DEBUG": "true"
}
}
Run Diagnostics
diagnostic() // Available in debug mode
Common Issues
- Missing API Keys: Ensure all required environment variables are set.
- ffmpeg Not Found: Install ffmpeg using
brew install ffmpeg. - Permission Issues: Check file system permissions for project directories.
Contributing
We welcome contributions! Please see our Contributing Guide for details.
๐ License
MIT ยฉ Data Route LLC
Support
- ๐ Report Issues
- ๐ฌ Support
BolideAI MCP - Streamline your marketing automation and research workflows with AI - powered tools.








