Beta
Flow is in beta. We are actively shaping the product, so things may change as we iterate. Your feedback is welcome.
Flow Overview
Flow is Mia-Platform's AI-driven development web application. You open it in the browser, describe what you want to build in natural language, and the assistant generates the code, runs it live in a sandboxed preview, and deploys it to a Mia-Platform project when you are ready.
Flow brings together three things developers normally juggle separately:
- A chat-first interface for talking to an AI assistant that understands your project and the systems you have connected.
- A live preview of the generated code, updated as you iterate with the assistant.
- A catalog of reusable AI building blocks (agents, skills, prompts, specs, playbooks) that pre-configure the assistant for specific tasks, teams, or domains.
What you can do with Flow
- Generate full-stack applications from a prompt. Frontend, backend, database schema, authentication, tests, and documentation are produced from a single natural-language description and iterated through follow-up messages.
- Preview the result instantly. Each generation is rendered in a live preview inside the Canvas, so you can interact with the application as it grows.
- Reach external systems through MCP. The assistant can call tools exposed by GitHub, GitLab, Azure DevOps, Atlassian, Grafana, Google Drive, Mia-Platform itself, and any custom MCP server you register.
- Deploy to Mia-Platform. A connected Mia-Platform project becomes the target for the generated code: pushing from Flow triggers the project's CI/CD pipeline.
The main areas of the application
| Area | What it is for |
|---|---|
| Home | Quick-start omnibar and recent conversations. |
| Chat | Conversational interface with the AI assistant. |
| Canvas | Code editor and live preview for the project tied to a conversation. |
| AI Foundry | Manage reusable agents, skills, prompts, specs, and playbooks. |
| Connectors | Authenticate to the external systems Flow can act on. |
| Memories | Browse, rename, and reopen past conversations. |
| Settings | Defaults, advanced options, and preferences. |
Where to go next
- Getting started: sign in and open your first conversation.
- Connected tools: understand how MCP integrations expose external systems to the AI.
- Chat: the conversational interface and how sessions are persisted.
- Code: the Canvas, live preview, and supported frameworks.
- Agentic AI: agents, skills, playbooks, and how the assistant reasons.