Can I publish a workflow immediately after creating it?

No. Drafts must be validated with prepare_publish and then explicitly confirmed with confirm_publish.

Does the server require database credentials?

No. It accesses only workflow metadata and analytics through the Dreamlit platform.

Which scopes should I request for read-only access?

Request only workflows:read for listing and inspecting workflows and brand kits.

Dreamlit

Verified

Create, test, publish, and manage Dreamlit notification workflows from AI clients.

MCP ServerAI & KnowledgeRemote (streamable-http)

View on GitHub

Updated 2026-06-16

What is the Dreamlit MCP server?

Dreamlit MCP lets AI clients build notification workflows using natural language prompts that specify events, audiences, goals, and timing. It provides tools for listing projects and workflows, applying brand kits, previewing drafts, sending tests, and handling publish or unpublish actions with explicit confirmation steps.

Analytics queries are bounded and paginated for recipient engagement and run data. The server accesses only workflow metadata and connected project context without requiring clients to transmit database credentials.

Install & connect

Add this to your MCP client config. Pick your client below and copy.

{
  "mcpServers": {
    "mcp": {
      "url": "https://mcp.dreamlit.ai/mcp"
    }
  }
}

Tools it exposes

11 tools your AI client can call once connected.

get_status— Get Dreamlit guidance, workspace context, project setup, schema hints, workflow state, and app URLs.

list_projects— Find accessible Dreamlit projects in the approved workspace.

list_workflows— Find workflows in a project by name, description, trigger type, and pagination.

list_brand_kits— Find saved brand kit ids and style summaries with pagination before applying one to generated email drafts.

get_workflow_and_preview_url— Inspect a workflow draft and get the Dreamlit preview or builder URL before editing or publishing.

get_analytics— Query bounded notification analytics, recipient engagement, and workflow run data with filters and cursor pagination.

create_or_update_workflow— Create a new workflow draft or update an existing draft from a natural-language prompt.

send_workflow_test— Confirm and send a draft email or Slack test without publishing the workflow.

prepare_publish— Validate a workflow before publishing and return confirmation fields.

confirm_publish— Publish or schedule a workflow after explicit user confirmation.

unpublish_workflow— Disable live triggers or schedules for a workflow.

Example prompts

Once connected, try asking your AI client:

“When a new organization row is inserted, send an internal Slack alert with the organization name, owner email, and plan.

“Every Monday at 9 AM America/New_York, send admins a weekly usage digest summarizing active users, failed payments, and new signups.

“Tomorrow at 10 AM, announce the launch to active users who opted into product updates. Include unsubscribe.

Security & permissions

Remote Streamable HTTP server at https://mcp.dreamlit.ai/mcp using OAuth (or personal access token fallback) scoped to a selected workspace. It touches only workflow metadata, brand kits, and paginated analytics results.

What you can do with Dreamlit

Workflow creation from prompts

Describe an event, audience, and message goal to generate or update notification drafts for email or Slack.

Pre-publish validation and testing

Inspect drafts, apply brand styles, send test messages, and validate before scheduling or publishing.

Bounded analytics review

Query engagement metrics and workflow performance using filters and cursors without exposing raw database access.

How to use Dreamlit

1Add the server in your MCP client using URL https://mcp.dreamlit.ai/mcp and complete OAuth login with desired scopes.
2Call get_status first to retrieve workspace context and prompting guidance.
3Use list_projects or list_workflows to locate resources when IDs are unknown.
4Create or update drafts with create_or_update_workflow then preview and test before publishing.
5Always use prepare_publish followed by confirm_publish for any live deployment.

Dreamlit: pros & cons

Pros

+Outcome-focused natural language prompting reduces need for manual graph editing.
+Explicit confirmation steps and scoped OAuth minimize accidental publishes.
+Built-in analytics and brand-kit tools keep workflows consistent without extra APIs.
+Remote hosted server requires no local installation or database credential sharing.

Cons

–Backend implementation is not open source.
–Publishing always requires separate multi-step confirmation.
–Analytics queries are intentionally bounded and may need multiple paginated calls.

Did you find this helpful?

Frequently asked questions

It uses Streamable HTTP transport with OAuth; personal access tokens are available as fallback.

User reviews

Verified reviews from the community shape this listing's rating.

Loading reviews…

Promote Dreamlit