Bolide AI MCP Server

How to use

You run the MCP server locally or connect to it from an MCP client. You can scaffold projects, perform research, generate content, manage Linear issues, automate Slack and Notion tasks, and run diagnostic checks. Start by configuring one or more MCP servers, then invoke tools from your client as needed. Enable the appropriate tool groups to optimize performance, and use environment variables to secure API keys and tokens.

When you start an MCP server in stdio mode, you run a local command that launches the server process and exposes its tools to your MCP client. Your client can request operations like scaffolding a project, performing research with Perplexity/OpenAI, analyzing screencasts, or integrating with Notion, Slack, and Linear. Use the supplied environment variables to authorize access to services and enable diagnostic logging for debugging.

How to install

# Prerequisites
# 1) Install Node.js 22+
# 2) Install FFmpeg if you plan to generate GIFs

# Quick Start: install MCP globally and run via npx
npm install -g @bolide-ai/mcp

# Or run directly with npx (recommended)
npx --package=@bolide-ai/mcp bolide-ai-mcp

# Optional: set up a local development MCP server (see Development Setup)

To run a local, development-oriented MCP server, you can follow the development steps to build and run the server from source. This typically involves cloning the repository, installing dependencies, building the server, and then starting it with a stdio command such as node build/index.js.

Configuration and usage notes

Configure your MCP client to connect to one or more servers. The primary configuration shows a stdio server that runs via npx or node, and environment variables provide access tokens for API services.

Common environment variables you might set include BOLIDE_AI_API_KEY for research tools and BOLIDE_AI_API_TOKEN for token-based access in development workflows. You can also enable BOLIDE_AI_MCP_DEBUG to activate diagnostic tools and verbose logging during troubleshooting.

Troubleshooting

If you encounter issues, enable debug mode to surface detailed logs and run diagnostics to validate the environment. Check for missing API keys, ensure ffmpeg is installed for media tooling, and verify file system permissions for project directories.

scaffold_bolide_ai_project

Create bolide.ai project directory structure for new campaigns or projects

check_companion_app_status

Check whether the companion app is running and reachable

launch_companion_app

Launch the companion app to capture marketing content such as screenshots and videos

stop_companion_app

Terminate running instances of the companion app

install_brew_and_ffmpeg

Install Homebrew and FFmpeg for media processing and GIF creation

analyze_screencasts

Analyze screencast content using Gemini AI to extract insights

generate_gif

Convert screencast segments into GIFs for sharing

enhance_audio

Extract and enhance audio from screencasts using ElevenLabs

use_perplexity

Research using Perplexity AI for quick information gathering

use_openai_deep_research

Deep research using OpenAI o4-mini-deep-research with web access

slack_fetch_conversation_history

Fetch chronological message history from Slack conversations

slack_list_all_slack_team_channels

List all channels, DMs, and group messages in the workspace

slack_search_for_messages_with_query

Search messages across workspace with query modifiers

slack_sends_a_message_to_a_slack_channel

Send messages to channels, DMs, or groups

slack_updates_a_slack_message

Update existing Slack messages by channel and timestamp

notion_add_page_content

Add content blocks to Notion pages with rich formatting

notion_fetch_data

Fetch pages and databases from Notion workspace

notion_create_comment

Add comments to Notion pages or discussion threads

notion_create_database

Create new Notion databases with defined property schemas

notion_create_notion_page

Create new empty Notion pages

notion_fetch_database

Fetch database structure and metadata

notion_fetch_row

Retrieve database row properties and metadata

notion_insert_row_database

Create new pages (rows) in specified databases

notion_query_database

Query database for pages (rows) with sorting and pagination

notion_retrieve_database_property

Get details about specific database properties

notion_update_page

Update Notion page properties, icon, cover, or archive status

notion_update_row_database

Update or archive existing database rows

notion_update_schema_database

Update database title, description, and properties

notion_append_block_children

Append child blocks to parent blocks or pages

notion_fetch_notion_block

Retrieve specific blocks by UUID

notion_fetch_notion_child_block

Get paginated list of child blocks

notion_notion_update_block

Update block content and type-specific properties

notion_search_notion_page

Search pages and databases by title

linear_create_issue

Create new Linear issues with comprehensive options

linear_update_issue

Update existing Linear issues (title, description, state, assignee, etc.)

linear_create_comment

Add comments to Linear issues

linear_list_issues

List Linear issues with filtering and pagination

linear_list_cycles

Get all cycles/sprints from Linear

linear_get_cycles_by_team_id

Get team-specific cycles

linear_list_states

Get workflow states for teams

linear_list_teams

Get teams with project filtering

linear_list_projects

List all Linear projects

linear_list_users

List workspace users with pagination

diagnostic

System environment validation (debug mode only)