Tianzeyui/stardust — Open-source AI Agent platform with React/TSX & HTML plugin dev modes.

Description and use

**Brief Summary**
Tianzeyui / stardust is an open‑source AI‑agent platform that lets developers build, run and extend intelligent assistants using a React/TSX front‑end and an HTML‑based plugin system. It ships with more than 60 ready‑to‑use APIs, multi‑project management, MCP (Model Context Protocol) tooling and a sandbox execution engine, creating a shared ecosystem for both official and community‑driven agents.

**Value**

• **Standardised tool integration** – By exposing a common protocol (MCP) and a rich set of pre‑built APIs, Stardust removes the friction of wiring large language models to real‑world services, data stores, and UI components.
• **Extensible plug‑in model** – The React/TSX + HTML plug‑in architecture lets teams add custom UI widgets or backend adapters without touching the core engine, fostering rapid experimentation and reuse across projects.
• **Unified development experience** – A single codebase provides both front‑end (React) and back‑end (TypeScript) tooling, plus CLI/SDK helpers, so engineers can develop, test and ship agents end‑to‑end in one workflow.

**Practical Adoption Path**

1. **Evaluate the SDK/CLI** – Clone the repo, run the provided CLI to spin up a local sandbox, and explore the 60+ built‑in API definitions.
2. **Prototype a connector** – Use the TypeScript SDK to wrap an existing internal service (e.g., a CRM API) as a new MCP endpoint; the plug‑in system lets you expose it via a simple HTML UI component.
3. **Deploy a Model Context Protocol server** – Package the agent as a MCP server (Docker or serverless) and point your LLM provider (OpenAI, Anthropic, etc.) to it, enabling the model to call your custom tools in production.
4. **Scale & share** – Leverage the multi‑project workspace to manage several agents, publish plug‑ins to the community hub, and integrate with CI/CD pipelines for continuous delivery.

**Production Readiness**

• **Activity & community** – The project was updated today (2026‑06‑24), has 41 stars, and shows recent commits, indicating active maintenance.
• **Technical maturity** – Built in TypeScript with a clear API surface, CLI, and sandbox engine; the codebase is modular and includes both front‑end and back‑end components, which simplifies testing and CI integration.
• **Ecosystem fit** – The MCP focus aligns with emerging standards for LLM tool usage, making it a future‑proof choice for enterprises looking to lock in a vendor‑agnostic integration layer.
• **Remaining checks** – Before a full production rollout, verify the open‑source license compatibility, conduct a security audit of the sandbox isolation, and confirm that a dedicated maintainer is assigned to respond to issues.

Overall, Stardust is a high‑readiness OSS candidate for organizations that need a standardized, extensible bridge between AI assistants and their internal tooling ecosystem.

Problem

Tianzeyui/stardust helps connect AI assistants to real tools and data through a standard protocol.

Outcome

High for an OSS candidate: recent activity, adoption, and ecosystem signals are strong enough for a serious pilot.

Workflow

• connect AI agents to tools
• ship Model Context Protocol servers
• standardize integrations

Evidence

Open-source AI Agent platform with React/TSX & HTML plugin dev modes. Built-in 60+ APIs, multi-project, MCP tools, sandbox engine. Official + community-driven, building an extensible agent ecosystem for everyone.

Source