Kasal turns complex AI orchestration into an intuitive visual experience. Design, deploy, and monitor autonomous AI agents that collaborate to solve real-world business problems — without writing orchestration code.
- Visual Workflow Designer — Drag-and-drop canvas for composing sophisticated agent interactions
- Enterprise-Ready — Built for Databricks with OAuth, teamspace isolation, and scale in mind
- Extensible Toolkit — A rich library of tools, including Genie, MCP servers, custom APIs, and data connectors
- Real-Time Monitoring — Live execution tracking with detailed logs, traces, and performance insights
- Production-Grade — Robust error handling, retry logic, and enterprise deployment patterns
- Data Analysis Pipelines — Agents that query, analyze, and visualize your data
- Content Generation Systems — Collaborative agents for research, writing, and content creation
- Business Process Automation — Intelligent workflows that adapt and make decisions
- Customer Support Assistants — Multi-agent systems with specialized knowledge domains
- Research & Development — Agents that gather, synthesize, and present insights
Install directly from the Databricks Apps Marketplace with one click — the best path for production, with automatic updates and enterprise support.
Use the deployment script in this repository for custom installations. Ideal for tailored configurations and advanced setups.
A quick setup for testing and development — requires Python 3.11 (pinned >=3.11,<3.12), uv and Node.js 22.
cd src/backend && ./run.sh # API on http://127.0.0.1:8000, SQLite by default
cd src/frontend && npm ci && npm start # UI on http://localhost:3000run.sh syncs dependencies with uv, binds to loopback, and enables a local development identity (LOCAL_DEV_AUTH). See the Quick Start for details.
The visual workflow designer for building AI agent collaborations
Create your first agent workflow in under two minutes:
- Design — Drag agents onto the canvas and define their roles
- Connect — Link agents together to form collaboration flows
- Execute — Hit run and watch your agents work as a team
- Monitor — Follow real-time logs and execution traces
| Topic | Description |
|---|---|
| Why Kasal | What problems it solves and who it's for |
| Solution Architecture | Layers, lifecycles, and platform integration |
| Code Structure | Where things live and how to navigate the repo |
| Developer Guide | Local setup, config, and extension patterns |
| API Reference | REST endpoints, payloads, and errors |
| Configuration | Environment variables, defaults, and where they are read |
- Docs Hub - Documentation index
- End‑User Tutorial Catalog - Screenshot-ready walkthroughs
- Testing Guide - Testing strategy and coverage
- Continuous Integration - CI workflows and how to run the checks locally
- Contributing - How to set up, test, and submit a change
Kasal follows a clean, layered architecture designed for scalability and maintainability:
Frontend (React) → API (FastAPI) → Services → Repositories → Database
Agents run on Kasal's own runtime (src/backend/src/services/execution/runtime/), with CrewAI available as an alternative harness. See Harnesses.
Licensed under the Databricks License
Unlocking Databricks Marketplace: A Hands-On Guide for Data Consumers
