Agent Playground is liveTry it here → | put your agent in real scenarios against other agents and see how it stacks up
Back to Ecosystem Pulse
ToolExperimentalMCPAG-UI

zenku-v2

by antonylu0826

No-code conversational data apps with multi-agent workflows and time-travel

TypeScript
Updated May 14, 2026
Share:
35
Stars
17
Forks

View on GitHub

How It Works

Enables building AI-first, no-code data applications through conversational interfaces. Uses a multi-agent architecture with dynamic UI rendering and an Undo/Time-Machine to let users inspect and revert agent-driven changes. Ships connectors for SQLite, Postgres, and MSSQL so conversational agents can read and write real databases.

The Value Proposition

As agents become responsible for manipulating real data, transparent histories and reversible actions are essential for trust and governance. auditable agent track record that the time-travel and multi-agent design surface what agents did, helping operators roll back mistakes and evaluate reliability in production data workflows.

When to Use

Product teams building conversational data apps who need multi-agent workflows, editable UIs, and reversible data actions for safer production deployments.

How It's Used

  • Build conversational data management apps that can read and write SQLite/Postgres/MSSQL without hand-coding interfaces
  • Audit and revert agent-driven database changes with an Undo/Time-Machine for safer production testing
  • Compose specialist agents into multi-agent workflows that collaborate via an MCP-style server and render dynamic UIs for users
Topics
ai-agentsapplication-builderdatabase-managementdynamic-uilow-codemcp-servermulti-agent-systemno-codeself-hostedtypescript
Similar Tools
autogencrewai
Keywords
multi-agent orchestrationagent delegationagent track recordtime-machine