Provides AI-powered search and integrations across GitLab, Jira, Confluence, YouTube, and more for automated workflows.
Configuration
View docs{
"mcpServers": {
"athapong-aio-mcp": {
"command": "aio-mcp",
"args": [
"-env",
"/path/to/.env",
"-sse",
"-sse-addr",
":8080",
"-sse-base-path",
"/mcp"
]
}
}
}You run this MCP server to access AI-powered tools and data from multiple sources through a single, unified interface. It enables real-time communication with clients, supports a suite of integrations, and powers advanced reasoning and search across connected services.
You interact with the MCP server using a client that supports either Standard I/O (stdio) or Server-Sent Events (SSE). In stdio mode, the server communicates via standard input/output, which is the default, suitable for desktop clients. In SSE mode, the server runs as an HTTP server, exposing endpoints for real-time event streaming and message exchange. Enable SSE when you need network access or web-based clients.
Prerequisites: install Go 1.23.2 or higher and prepare API keys/tokens for the services you intend to use.
Install via Smithery (client setup is interactive):npx -y @smithery/cli install @athapong/aio-mcp --client claude
Install via Go (build and run the MCP server locally):1) Set up your Go environment so binaries go to your GOPATH/bin and that directory is on your PATH. 2) Install the server BINARY:``
go install github.com/athapong/aio-mcp@latest
``
Create a configuration file (.env) with your credentials and options. Here is a representative set of environment variables you may configure: ENABLE_TOOLS, QDRANT_HOST, ATLASSIAN_HOST, ATLASSIAN_EMAIL, GITLAB_HOST, GITLAB_TOKEN, BRAVE_API_KEY, ATLASSIAN_TOKEN, GOOGLE_AI_API_KEY, OPENAI_API_KEY, OPENAI_EMBEDDING_MODEL, PROXY_URL, QDRANT_PORT, GOOGLE_TOKEN_FILE, GOOGLE_CREDENTIALS_FILE, QDRANT_API_KEY, USE_OLLAMA_DEEPSEEK, ENABLE_SSE, SSE_ADDR, SSE_BASE_PATH, and related keys for deep search and OpenRouter if used.
Configure the MCP server entry in your Claude’s config (example below). This uses a stdio runtime with environment file support and SSE enabled at port 8080 with a base path of /mcp.
To run in SSE mode, you can start the MCP server with command-line flags or via environment variables. Command-line: aio-mcp -sse -sse-addr ":8080" -sse-base-path "/mcp". Environment variables (in your .env): ENABLE_SSE=true, SSE_ADDR=:8080, SSE_BASE_PATH=/mcp.
Example of the MCP server initialization in a client config file is provided below to illustrate how the server is wired for stdio operation and environment-based configuration.
{
"mcpServers": {
"aio_mcp": {
"type": "stdio",
"command": "aio-mcp",
"args": ["-env", "/path/to/.env", "-sse", "-sse-addr", ":8080", "-sse-base-path", "/mcp"]
}
}
}Keep your API keys and tokens secure. Do not commit the .env file to version control. Use appropriate file permissions and consider using a secret manager for production deployments.
Only enable tools you trust. The server supports a range of tool groups (e.g., gemini, fetch, brave_search, google_maps, confluence, youtube, jira, gitlab, script, rag, deepseek). You can selectively enable or disable tool groups via environment controls.
If SSE mode does not start, verify that the chosen address is not in use and that the .env file is accessible by the runtime. Check logs for authentication errors related to external services. Ensure the CLI client and server packages are compatible with your Go version.
Create a new event in Google Calendar with title, description, start/end times, and attendees.
List upcoming Google Calendar events within a specified time range.
Update details of an existing Google Calendar event.
Respond to a Google Calendar event invitation.
Search content in Confluence using CQL.
Retrieve content of a Confluence page by ID.
Create a new Confluence page with specified space, title, and content.
Update an existing Confluence page with new content.
Compare two versions of a Confluence page.
Advanced multi-step reasoning using Deepseek AI capabilities.
Fetch and return content from a given HTTP/HTTPS URL.
List all Google Chat spaces available to you.
Send a message to a Google Chat space or direct message.
Query the web via Google AI Search for up-to-date information.
List GitLab projects under a group.
Get details for a specific GitLab project.
List merge requests for a GitLab project.
Retrieve details for a specific GitLab merge request.
Add a note to a GitLab merge request.
Fetch a file's content from a GitLab repository.
List pipelines for a GitLab project.
List commits for a GitLab project within a date range.
Get details of a specific commit in a GitLab project.
List user events in GitLab within a date range.
List all users in a GitLab group.
Create a new GitLab merge request.
Clone or update a GitLab repository locally.
Search Gmail messages using Gmail search syntax.
Move specified Gmail messages to spam.
Create a Gmail filter with criteria and actions.
List all Gmail filters in the account.
List all Gmail labels in the account.
Delete a Gmail filter by its ID.
Delete a Gmail label by its ID.
Retrieve detailed Jira issue information.
Search Jira issues using JQL.
List active and future sprints for a Jira board.
Create a new Jira issue with specified details.
Update an existing Jira issue with new fields.
Retrieve all available Jira issue statuses for a project.
Transition a Jira issue through its workflow.
Index content into an in-memory vector store.
Index a local file into memory as a vector.
Create a new in-memory vector collection.
Delete a vector collection from memory.
List all in-memory vector collections.
Search for memory in a collection using a query.
Delete a vector index by file path.
Safely execute command-line scripts with sandboxing and safety checks.
Search the web using Brave Search API.
Dynamic multi-step thinking tool to plan, revise, and verify solutions step-by-step.
Fetch YouTube video transcripts.
Update a YouTube video's title and description.
Get details for a YouTube video.
List YouTube videos managed by the user.