PipesHub — Claude Code MCP server: what it does, how to install
Searches internal workspace platforms like Slack, Jira, and Google Drive while enforcing user permissions. Helpful when agents need direct context from team tools.
claude mcp add --env PIPESHUB_MCP_CHAT=<PIPESHUB_MCP_CHAT> --transport stdio pipeshub -- npx -y @pipeshub-ai/mcpWhat PipesHub does
PipesHub connects AI assistants to internal company knowledge bases. It provides MCP tools like pipeshub_search and pipeshub_get_record_content to locate entries and inspect text. Other available actions handle file downloads via pipeshub_download_record, directory discovery, data sources listing, and direct chat interactions.
Clients connect over Streamable HTTP at the /mcp path or via stdio execution. Authentication relies on bearer tokens or static OAuth with permission scopes. Self-hosted instances return HTTP 500 errors on search operations until an LLM is configured in the web interface.
What it helps with
- retrieve relevant Jira tickets without leaving the code editor
- search discussions in Slack channels under existing user credentials
- read document contents stored in Google Drive and Gmail
- inspect available corporate data connectors via MCP tools
Example prompts for the agent
- Find Jira issues related to authentication bugs reported this month.
- Search Slack for mentions of the API migration deadline.
- Fetch the onboarding guide document from Google Drive.
How to install PipesHub
1Platform CLI
claude mcp add --env PIPESHUB_MCP_CHAT=<PIPESHUB_MCP_CHAT> --transport stdio pipeshub -- npx -y @pipeshub-ai/mcpReplace <…> with your own values
2Platform CLI: remote server
claude mcp add --transport http pipeshub https://{instance_host}/mcp1Platform CLI
codex mcp add pipeshub --env PIPESHUB_MCP_CHAT=<PIPESHUB_MCP_CHAT> -- npx -y @pipeshub-ai/mcpReplace <…> with your own values
2Platform CLI: remote server
codex mcp add pipeshub --url https://{instance_host}/mcp1Config file
{
"mcpServers": {
"pipeshub": {
"command": "npx",
"args": [
"-y",
"@pipeshub-ai/mcp"
],
"env": {
"PIPESHUB_MCP_CHAT": "<PIPESHUB_MCP_CHAT>"
}
}
}
}Replace <…> with your own values
2Config file
{
"mcpServers": {
"pipeshub": {
"url": "https://{instance_host}/mcp"
}
}
}1Config file
{
"servers": {
"pipeshub": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"@pipeshub-ai/mcp"
],
"env": {
"PIPESHUB_MCP_CHAT": "<PIPESHUB_MCP_CHAT>"
}
}
}
}Replace <…> with your own values
2Config file
{
"servers": {
"pipeshub": {
"type": "http",
"url": "https://{instance_host}/mcp"
}
}
}1Platform CLI
gemini mcp add -e PIPESHUB_MCP_CHAT=<PIPESHUB_MCP_CHAT> pipeshub npx -y @pipeshub-ai/mcpReplace <…> with your own values
2Platform CLI: remote server
gemini mcp add -t http pipeshub https://{instance_host}/mcpStep by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about PipesHub
How to install PipesHub in Claude Code?
Run claude mcp add --transport http pipeshub https://{instance_host}/mcp to connect via HTTP. For a stdio setup, execute claude mcp add --env PIPESHUB_MCP_CHAT=<PIPESHUB_MCP_CHAT> --transport stdio pipeshub -- npx -y @pipeshub-ai/mcp.
What does the PipesHub MCP server do?
It gives coding agents read access to company platforms like Jira, Slack, and Google Drive. Retrieved content respects access permissions assigned to the user.
Why does PipesHub return a 500 error during search?
Search fails with an HTTP 500 status if an LLM has not been connected to the instance yet. Initial account creation and model selection must be finished in the web UI.
