Bubble Lab: An Open-Core Workflow Engine for AI-Powered Automation

Introduction to Bubble Lab

Bubble Lab represents a groundbreaking advancement in workflow automation, particularly within the Slack ecosystem. It is an open-core platform designed to empower teams with seamless integration of artificial intelligence (AI) and operational workflows directly inside Slack. Unlike traditional workflow management systems that require switching between multiple tools, Bubble Lab enables users to execute tasks, access databases, and interact with APIs—all within a unified interface known as Pearl, its AI-powered assistant.

This repository houses the open-core workflow engine that underpins Bubble Lab’s platform. It is fully functional, production-ready, and can be independently hosted, extended, or customized for various use cases. Whether you are part of an organization leveraging Bubble Lab’s managed services or a developer seeking full control over workflow execution, this engine provides the necessary infrastructure to build, deploy, and manage automated processes.

Core Features and Capabilities

1. Open-Core Architecture

Bubble Lab follows an open-core model, allowing users to choose between two primary deployment options:

Option 1: Fully Managed Bubble Lab Platform (Recommended)

For teams that prefer a managed solution, Bubble Lab offers a fully hosted platform with:

Pearl: A Slack-native AI assistant that simplifies workflow execution.
Managed Integrations: Seamless connections to Slack, SaaS tools, APIs, and databases.
Workflow Execution & Orchestration: Robust infrastructure for running automated processes.
Observability Dashboards: Detailed tracking of workflow execution history and performance metrics.
Team Collaboration Tools: Facilitates deployment management and real-time collaboration.

Users can access the platform via app.bubblelab.ai.

Option 2: Self-Hosted Open-Core Engine

For developers, engineers, or organizations requiring full control over their automation infrastructure, Bubble Lab’s open-core engine allows:

Local Workflow Development: Build and execute workflows directly on your machine.
Custom Agent & Integration Creation: Develop bespoke agents and integrations tailored to specific needs.
Self-Hosted Deployment: Run the engine in any infrastructure environment.
Workflow Export & Embedding: Deploy Bubble Lab workflows into custom applications or embed them within existing products.

2. Key Components of the Open-Core Engine

The repository provides a comprehensive set of tools and primitives essential for workflow automation:

A. Workflow Execution Runtime

The engine supports real-time execution of workflows, ensuring efficient processing with minimal latency. It includes:

Bubble Framework: A modular system where individual tasks (called "bubbles") are executed sequentially or in parallel.
Error Handling & Retry Mechanisms: Robust error management to handle failures gracefully.

B. Agent and Integration Primitives ("Bubbles")

"Bubbles" are the building blocks of workflows, representing tools or actions that can be chained together:

Webhook/HTTP Bubbles: Allow interaction with external APIs.
Database Bubbles: Enable database queries and updates.
AI Agent Bubbles: Integrate AI models for natural language processing (NLP) tasks.

C. Local Workflow Studio

A user-friendly interface for designing, editing, and previewing workflows without needing a hosted environment.

D. Execution Tracing & Observability

Comprehensive logging and metrics tracking to monitor workflow performance:

Execution History: Detailed logs of each step in a workflow.
Token Usage Analytics: Tracks API calls and AI model usage for cost optimization.
Memory & Performance Metrics: Ensures efficient resource utilization.

E. CLI Tooling

Command-line interfaces simplify deployment, debugging, and management:

# Install dependencies
pnpm install

# Start the local development server
pnpm run dev

This setup allows users to launch Bubble Lab Studio at http://localhost:3000 for immediate workflow testing.

Getting Started with Bubble Lab

Option A: Using the Managed Platform

No setup is required. Users can immediately start building and deploying workflows via:

Option B: Running Locally

For developers who prefer self-hosted workflows, the following steps outline how to run Bubble Lab locally:

Step 1: Install Dependencies

pnpm install

This installs all necessary packages, including TypeScript and runtime dependencies.

Step 2: Start Development Server

pnpm run dev

After execution, the local server will launch at http://localhost:3000, enabling workflow editing and testing.

Note: To use Pearl (the AI assistant), an API key is required. By default, the engine uses Google’s Gemini-3.0-Pro, which provides high accuracy. For alternative models, users can specify a weaker model, though performance may vary.

Option C: Creating a New Bubble Lab Project

For those who want to scaffold a new project quickly:

npx create-bubblelab-app

This command generates:

A preconfigured TypeScript setup with core packages.
Sample templates (e.g., basic workflows, Reddit scraper).
All necessary dependencies and ready-to-run examples.

Example: The Reddit Scraper Workflow

The repository includes a well-documented example workflow named reddit-scraper, which demonstrates how to:

Fetch posts from a specified Reddit subreddit.
Use AI to summarize the content.
Return structured JSON with metadata and summaries.

Here’s a snippet of the workflow code (reddit-news-flow.ts):

export class RedditNewsFlow extends BubbleFlow<'webhook/http'> {
  async handle(payload: RedditNewsPayload) {
    const subreddit = payload.subreddit || 'worldnews';
    const limit = payload.limit || 10;

    // Step 1: Scrape Reddit for posts
    const scrapeResult = await new RedditScrapeTool({
      subreddit,
      sort: 'hot',
      limit,
    }).action();
    const posts = scrapeResult.data.posts;

    // Step 2: AI analyzes and summarizes the posts
    const summaryResult = await new AIAgentBubble({
      message: `Analyze these top ${posts.length} posts from r/${subreddit}: ${postsText}
               Provide: 1) Summary of top news, 2) Key themes, 3) Executive summary`,
      model: { model: 'google/gemini-2.5-flash' },
    }).action();

    return {
      subreddit,
      postsScraped: posts.length,
      summary: summaryResult.data?.response,
      status: 'success',
    };
  }
}

Execution Example

When executed, the workflow:

Scrapes 10 hot posts from r/worldnews.
Uses an AI agent (Google Gemini) to generate a structured summary.
Returns a JSON response with metadata and analysis.

Output:

{
  "subreddit": "worldnews",
  "postsScraped": 10,
  "summary": "### Top 5 News Items:\\n1\. China Halts US Soybean Imports...\\n2\. Zelensky Firm on Ukraine's EU Membership...\\n3\. Hamas Demands Release of Oct 7 Attackers...",
  "timestamp": "2025-10-07T21:35:19.882Z",
  "status": "success"
}

Execution Summary:

Total Duration: 13.8 seconds
Bubbles Executed: 3 (RedditScrapeTool → AIAgentBubble → Return)
Token Usage: 1,524 tokens (835 input, 689 output)
Memory Peak: 139.8 MB

Understanding the Underlying Workflow Execution

The workflow execution process involves several key steps:

RedditScrapeTool: Fetches and processes posts from a specified Reddit subreddit.
AIAgentBubble: Uses an AI model (Google Gemini) to analyze the scraped content, generating summaries, themes, and executive insights.
Return Statement: Compiles the results into structured JSON for further use.

Each bubble in the workflow is executed sequentially, with intermediate results passed between steps. The engine ensures proper error handling and performance tracking, providing detailed logs of execution.

Key Features Highlighted by Bubble Lab

1. Type-Safe Workflows

Bubble Lab leverages TypeScript for type safety, ensuring that all inputs and outputs adhere to predefined interfaces. This reduces errors during development and improves maintainability.

2. Simple & Intuitive Chaining

Workflows are built using a straightforward syntax:

new RedditScrapeTool().action()

This allows users to chain bubbles together effortlessly, creating complex workflows with minimal code.

3. Observable Execution

The engine provides real-time observability through detailed logs and metrics:

Execution History: Tracks each step of the workflow.
Performance Analytics: Monitors token usage, memory consumption, and latency.

4. Production-Ready Infrastructure

Bubble Lab’s open-core engine includes robust features for production deployment:

Error Handling: Graceful recovery from failures.
Metrics & Logging: Comprehensive tracking for debugging and optimization.
Scalability: Designed to handle high-volume workflows efficiently.

Documentation and Learning Resources

For users seeking in-depth guidance on building workflows, Bubble Lab offers extensive documentation:

BubbleLab Documentation: Covers node types, workflow tutorials, API references, and best practices.
Node-Specific Guides: Detailed explanations for each bubble type (e.g., webhook integrations, AI agents).
Workflow Building Tutorials: Step-by-step examples of creating custom workflows.

Community & Support

Bubble Lab fosters a collaborative environment through:

Discord Community: A space for discussions, bug reports, and feature requests.
GitHub Issues: Platform for reporting problems or suggesting enhancements.

Note: As of January 2026, the repository no longer accepts code contributions or pull requests. However, users are encouraged to:

Report bugs and provide feedback.
Share feature ideas in community discussions.
Improve documentation based on user suggestions.

License

The open-core components of Bubble Lab are licensed under Apache 2.0. This means:

Open-source contributions are allowed under the license terms.
The platform’s proprietary components (Pearl, hosted infrastructure) remain separate and not included in this repository.

Conclusion

Bubble Lab represents a transformative approach to workflow automation, combining open-core flexibility with powerful AI capabilities. Whether you’re part of an organization leveraging its managed services or a developer building custom agents, Bubble Lab provides the tools needed to streamline operations within Slack and beyond.

By offering both a fully managed platform and a self-hostable engine, Bubble Lab empowers teams to automate tasks efficiently while maintaining full control over their workflows. The repository’s open-source nature ensures transparency, allowing users to inspect, extend, and adapt the engine for diverse use cases—from simple data processing to complex AI-driven automation.

For those eager to explore further, starting with the provided templates (such as the Reddit scraper) offers a practical introduction to building workflows with Bubble Lab. The extensive documentation and community support ensure that users can quickly master the platform’s capabilities and contribute to its evolution.

Images from Input:

Bubble Lab Logo: (Note: This is a placeholder for the actual logo image, which would typically appear in the app interface.)
Editing Workflow GIF:
Running Workflow GIF:

These visuals illustrate the user experience of editing and executing workflows within Bubble Lab’s studio environment.

BubbleLab: Open-Core Workflow Engine