Vue Quick Start
Import the renderer CSS once from your app entry or CSS pipeline:
// main.ts
import 'markstream-vue/index.css'Then render Markdown:
<script setup lang="ts">
import MarkdownRender from 'markstream-vue'
type MarkdownRenderProps = InstanceType<typeof MarkdownRender>['$props']
const md: MarkdownRenderProps['content'] = `# Hello World\n\nThis is **bold** and this is *italic*.`
const customId: MarkdownRenderProps['customId'] = 'quick-start'
const isDark: MarkdownRenderProps['isDark'] = false
</script>
<template>
<MarkdownRender
:content="md"
:custom-id="customId"
:is-dark="isDark"
/>
</template>If you want prop-level hover, start with MarkdownRenderProps['content'], MarkdownRenderProps['customId'], MarkdownRenderProps['isDark'], or the matching template attributes above. Hovering the component name itself is usually less informative in Vue snippets.
Note: the packaged CSS is scoped under an internal .markstream-vue container to reduce global style conflicts. The root JavaScript entry does not inject this stylesheet, so keep the CSS import in your app shell. Use @import 'markstream-vue/index.css' layer(components); when your CSS pipeline needs precise layer ordering.
For dark theme variables, either add .dark on an ancestor or pass :is-dark="true" to scope dark mode to the renderer.
Optional: wrap with <client-only> for Nuxt/SSR.
See /nuxt-ssr for Nuxt-specific instructions.
Try this quickly in your app:
<script setup lang="ts">
import MarkdownRender from 'markstream-vue'
type MarkdownRenderProps = InstanceType<typeof MarkdownRender>['$props']
const md: MarkdownRenderProps['content'] = `# Hello world\n\nTry a simple Mermaid:\n\n\`\`\`mermaid\ngraph LR\nA-->B\n\`\`\`\n\nTry a simple D2:\n\n\`\`\`d2\ndirection: right\nClient -> API: request\nAPI -> DB: query\nDB -> API: rows\nAPI -> Client: response\n\`\`\`\n`
</script>
<template>
<MarkdownRender :content="md" />
</template>Install mermaid or @terrastruct/d2 to render those diagrams; without them the renderer falls back to showing source text.
For chat-style streaming, start with mode="chat" plus content; max-live-nodes <= 0 in chat mode enables smooth-streaming="auto" and keeps fade off by default. If another layer already parses in a worker/store or you need AST control, pass :nodes + :final instead. See /guide/usage, /guide/ai-chat-streaming, and /guide/performance.