Dreamlit — MCP server for Claude Code: features and setup
Connect an AI coding client to Dreamlit to draft notification workflows, test email or Slack messages, control publishing, and query recipient engagement.
claude mcp add --transport http mcp https://mcp.dreamlit.ai/mcpWhat Dreamlit does
Dreamlit connects AI clients to notification workflows for Supabase and Postgres applications. Its MCP tools let you browse projects and workflows, generate or revise drafts from plain-language requests, inspect previews, and apply saved brand kits to generated email drafts. Analytics queries cover notification activity, recipient engagement, and workflow runs, with filters and cursor-based pagination.
Testing and publishing are separate operations. You can send a draft email or Slack message for review, validate a workflow, and then publish or schedule it after explicit confirmation. Draft creation does not activate a workflow, so a generated notification will not go live merely because the AI finished writing it. Existing workflows can also be unpublished to stop their live triggers or schedules.
The hosted server uses Streamable HTTP and OAuth, with project-scoped personal access tokens available for clients that cannot use OAuth. Permissions separate workflow reading, editing, publishing, and analytics. MCP access does not require database credentials, but it also does not provide CSV exports or bulk data dumps. For larger analytics results, use filters and the returned pagination cursors.
What it helps with
- Draft a notification workflow from an event and recipient description
- Locate a workflow in an accessible Dreamlit project
- Review email or Slack test messages before going live
- Schedule a workflow after validation and user approval
- Inspect recipient engagement with filtered analytics queries
Example prompts for the agent
- Create a Dreamlit draft that sends an internal Slack alert when a new organization is added, including its name, owner email, and plan.
- Find the existing weekly usage digest workflow, return its preview URL, and prepare an email test.
- Analyze recipient engagement for the selected workflow and paginate through the matching results.
How to install Dreamlit
Platform CLI: remote server
claude mcp add --transport http mcp https://mcp.dreamlit.ai/mcpPlatform CLI: remote server
codex mcp add dreamlit --url https://mcp.dreamlit.ai/mcpConfig file
{
"mcpServers": {
"dreamlit": {
"url": "https://mcp.dreamlit.ai/mcp"
}
}
}Config file
{
"servers": {
"dreamlit": {
"type": "http",
"url": "https://mcp.dreamlit.ai/mcp"
}
}
}Platform CLI: remote server
gemini mcp add -t http dreamlit https://mcp.dreamlit.ai/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 Dreamlit
How do I install Dreamlit in Claude Code?
Run claude mcp add --transport http dreamlit https://mcp.dreamlit.ai/mcp. Complete the Dreamlit authorization flow and select the workspace and permissions the client should receive.
What does Dreamlit do?
It gives an AI client tools to draft and inspect notification workflows, test email or Slack messages, manage publication, and query workflow analytics.
Does Dreamlit MCP require a personal access token?
OAuth is the standard authorization method. A project-scoped personal access token is an alternative for clients that do not support OAuth.
