Docs/add vercel deployment guide zh

[inoribea]

Summary

Add comprehensive Chinese (Simplified) documentation for deploying AIRI Stage Web on Vercel.

Contents

• 📖 Overview: Introduction to Vercel deployment
• ⚙️ Prerequisites: Required tools and accounts
• 🚀 Step-by-step deployment:
  1. Import project into Vercel
  2. Configure environment variables
  3. Trigger builds
  4. Optional local verification
• 🔧 Environment Variables: Complete reference table including:
  • Default provider configuration
  • API keys for various providers (OpenAI, Anthropic, Google, etc.)
  • WebSocket endpoint configuration
• 🧠 Memory System Configuration: Detailed setup for:
  • Short-term memory (Redis/Upstash/Vercel KV)
  • Long-term memory (Postgres+pgvector/Qdrant)
  • Embedding providers
• 🐛 Troubleshooting: Common issues and solutions

File Structure

docs/content/zh-Hans/docs/guides/deploy/vercel.md

Related

• Mirrors the existing English guide at docs/content/en/docs/guides/deploy/vercel.md
• Provides Chinese users with complete deployment documentation
• Improves accessibility for Chinese-speaking developers

[netlify]

✅ Deploy Preview for airi-docs ready!

Name	Link
🔨 Latest commit	e192419
🔍 Latest deploy log	https://app.netlify.com/projects/airi-docs/deploys/68e50ec57f9ceb0008a9c369
😎 Deploy Preview	https://deploy-preview-637--airi-docs.netlify.app
📱 Preview on mobile	Toggle QR Code... Use your smartphone camera to open QR code link.

---

To edit notification comments on pull requests, go to your Netlify project configuration.

[netlify]

✅ Deploy Preview for airi-vtuber ready!

Name	Link
🔨 Latest commit	e192419
🔍 Latest deploy log	https://app.netlify.com/projects/airi-vtuber/deploys/68e50ec5acb2da0008dd9819
😎 Deploy Preview	https://deploy-preview-637--airi-vtuber.netlify.app
📱 Preview on mobile	Toggle QR Code... Use your smartphone camera to open QR code link.

---

To edit notification comments on pull requests, go to your Netlify project configuration.

[gemini-code-assist]

Summary of Changes

Hello @inoribea, I'm Gemini Code Assist¹! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!

This pull request significantly enhances the AIRI Stage Web application by integrating a comprehensive memory system, offering both short-term conversation context persistence and long-term semantic memory storage using various backend technologies. It also introduces a detailed deployment guide for Vercel in Simplified Chinese, broadening the application's reach. Further improvements include a new UI for memory configuration, expanded environment variable support for flexible deployments, and advanced functionality for importing display models from URLs.

Highlights

• New Chinese Vercel Deployment Guide: A comprehensive guide for deploying AIRI Stage Web on Vercel has been added in Simplified Chinese, mirroring the existing English version to improve accessibility.
• Memory System Implementation: A new @proj-airi/memory-system package has been introduced, providing robust short-term (Redis/Upstash/Vercel KV) and long-term (Postgres/Qdrant with embeddings) memory capabilities for conversations.
• Memory API Endpoints & UI: The server-runtime now exposes API endpoints for memory management, including configuration, saving, searching, clearing, and exporting. Corresponding UI components have been added to stage-pages for easy configuration.
• Enhanced Environment Variable Integration: Extensive environment variables have been added to vite-env.d.ts, vite.config.ts, and turbo.json to support Vercel deployment and flexible memory system configuration.
• Font Preloading & Optimization: New font dependencies have been added, and a fonts.css file is introduced to preload locally bundled fonts, optimizing font loading and reducing reliance on external CDN requests.
• Model Import from URL: The UI now supports importing display models (VRM, Live2D) directly from a URL, including advanced handling for model3.json and VPM JSON formats, fetching and zipping related resources.

Ignored Files

• Ignored by pattern: packages/i18n/src/** (1)
  • packages/i18n/src/locales/zh-Hans/settings.yaml

Using Gemini Code Assist

The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.

Invoking Gemini

You can request assistance from Gemini at any point by creating a comment using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands on the current page.

Feature	Command	Description
Code Review	/gemini review	Performs a code review for the current pull request in its current state.
Pull Request Summary	/gemini summary	Provides a summary of the current pull request in its current state.
Comment	@gemini-code-assist	Responds in comments when explicitly tagged, both in pull request comments and review comments.
Help	/gemini help	Displays a list of available commands.

Customization

To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.

Limitations & Feedback

Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here.

You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension.

Footnotes

1. Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩

[gemini-code-assist]

Code Review

This pull request is quite large and introduces several major features beyond the Chinese Vercel deployment guide mentioned in the title. These include a comprehensive memory system, extensive environment variable handling, local font bundling, and URL-based model imports. In the future, please consider splitting such significant changes into smaller, more focused pull requests to facilitate easier and more thorough reviews. My feedback focuses on improving maintainability, robustness, and code clarity in the new additions.

[gemini-code-assist]

Using void zodCore to prevent tree-shaking is a known workaround, but it can be a bit of a code smell. It indicates that the bundler isn't correctly detecting the side effects or dependencies required by xsschema.

Have you considered a more explicit configuration-based approach? For example, in Vite, you could use optimizeDeps.include in vite.config.ts to ensure the necessary zod chunks are always included in the bundle.

// in vite.config.ts
export default defineConfig({
  optimizeDeps: {
    include: ['zod/v4/core'], // Or the specific chunk needed by xsschema
  },
  // ...
})

This would move the bundling hint from the application code into the build configuration, which is generally a cleaner separation of concerns.

[gemini-code-assist]

This section adds a lot of detailed environment variable documentation, which is also present in the new docs/content/en/docs/guides/deploy/vercel.md file. Duplicating this information across multiple README files can lead to maintenance issues, as any changes would need to be updated in several places. To improve maintainability, consider removing these detailed tables from the README files and instead providing a link to the main Vercel deployment guide. Since you've also added a Chinese version of the guide, you can link to the appropriate language version from each README.

[gemini-code-assist]

This section adds a lot of detailed environment variable documentation, which is also present in the new docs/content/en/docs/guides/deploy/vercel.md file. Duplicating this information across multiple README files can lead to maintenance issues, as any changes would need to be updated in several places. To improve maintainability, consider removing these detailed tables from the README files and instead providing a link to the main Vercel deployment guide. Since you've also added a Chinese version of the guide, you can link to the appropriate language version from each README.

[gemini-code-assist]

This section adds a lot of detailed environment variable documentation, which is also present in the new docs/content/en/docs/guides/deploy/vercel.md file. Duplicating this information across multiple README files can lead to maintenance issues, as any changes would need to be updated in several places. To improve maintainability, consider removing these detailed tables from the README files and instead providing a link to the main Vercel deployment guide. Since you've also added a Chinese version of the guide, you can link to the appropriate language version from each README.

[gemini-code-assist]

This block adds several new API endpoints for the memory system. There's some repeated logic for request validation (e.g., checking for sessionId or userId) and error handling across these handlers. To improve code reuse and maintainability, consider the following:

1. Centralized Validation: Extract the validation logic into a shared utility function or H3 middleware. This would reduce code duplication.
2. Consistent Error Responses: When validation fails, the handlers return a JSON error object but with a 200 OK status. It would be more idiomatic to use appropriate HTTP status codes, like 400 Bad Request. You can do this in H3 by using setResponseStatus or throwing an error with createError.

For example:

import { createError, eventHandler, readBody, setResponseStatus } from 'h3'

// Example of a more robust handler
router.post('/api/memory/clear', eventHandler(async (event) => {
  const body = await readBody(event) as { sessionId?: string }

  if (!body?.sessionId || typeof body.sessionId !== 'string') {
    throw createError({
      statusCode: 400,
      statusMessage: 'Bad Request',
      data: { success: false, error: 'sessionId is required' }
    })
  }

  await clearSessionMemory(body.sessionId)
  setResponseStatus(event, 204) // No Content for successful deletion
}))

[gemini-code-assist]

The normalizeShortTermProvider function silently falls back to local-redis if upstash-redis is selected but the required environment variables are missing. This implicit fallback could be confusing for users who expect their configuration to be respected or to receive an error if it's incomplete. It would be more transparent to either:

1. Log a warning message indicating that the configuration is incomplete and that the system is falling back to local-redis.
2. Throw an error if the configured provider is missing its required environment variables, forcing the user to provide a complete configuration.

This would make debugging configuration issues much easier.

[gemini-code-assist]

The loadModel3JsonAndResources function is a great feature for importing Live2D models from a model3.json URL. However, it relies on fetching numerous individual files from a base URL, which can be fragile. If any of the resource files (textures, motions, etc.) fail to download, the catch block inside the map only logs a warning, and the process continues. This could result in a partially incomplete or corrupted model ZIP file being created and cached, which might lead to runtime errors or visual glitches later.

To make this more robust, consider either:

• Failing the entire import if a critical resource (like the .moc3 file or textures) cannot be fetched.
• Implementing a retry mechanism for failed network requests.
• At the very least, providing more prominent feedback to the user that some model resources failed to load.

[gemini-code-assist]

The env and passThroughEnv arrays are defined here for the build task, but they seem to be identical to the globalEnv and globalPassThroughEnv arrays defined at the root of the configuration. Since these are global, they should apply to all tasks, making the task-specific definitions redundant. To simplify the configuration, you can remove the env and passThroughEnv from the build task and rely on the global definitions.

[nekomeowww]

Are you sure? This will bundle all the API secret into visible text as part of bundle for everyone to see it.

[nekomeowww]

Suggested change

	readonly OPENAI_API_KEY?: string
	readonly OPENAI_BASE_URL?: string
	readonly OPENAI_MODEL?: string
	readonly OPENAI_COMPATIBLE_API_KEY?: string
	readonly OPENAI_COMPATIBLE_BASE_URL?: string
	readonly OPENAI_COMPATIBLE_MODEL?: string
	readonly OPENAI_SPEECH_MODEL?: string
	readonly OPENAI_COMPATIBLE_SPEECH_MODEL?: string
	readonly OPENAI_TRANSCRIPTION_MODEL?: string
	readonly OPENAI_COMPATIBLE_TRANSCRIPTION_MODEL?: string
	readonly OPENROUTER_API_KEY?: string
	readonly OPENROUTER_BASE_URL?: string
	readonly OPENROUTER_MODEL?: string
	readonly ANTHROPIC_API_KEY?: string
	readonly ANTHROPIC_BASE_URL?: string
	readonly ANTHROPIC_MODEL?: string
	readonly GOOGLE_GENERATIVE_AI_API_KEY?: string
	readonly GOOGLE_GENERATIVE_AI_BASE_URL?: string
	readonly GOOGLE_GENERATIVE_AI_MODEL?: string
	readonly DEEPSEEK_API_KEY?: string
	readonly DEEPSEEK_BASE_URL?: string
	readonly DEEPSEEK_MODEL?: string
	readonly AI302_API_KEY?: string
	readonly AI302_BASE_URL?: string
	readonly AI302_MODEL?: string
	readonly TOGETHER_API_KEY?: string
	readonly TOGETHER_BASE_URL?: string
	readonly TOGETHER_MODEL?: string
	readonly XAI_API_KEY?: string
	readonly XAI_BASE_URL?: string
	readonly XAI_MODEL?: string
	readonly NOVITA_API_KEY?: string
	readonly NOVITA_BASE_URL?: string
	readonly NOVITA_MODEL?: string
	readonly FIREWORKS_API_KEY?: string
	readonly FIREWORKS_BASE_URL?: string
	readonly FIREWORKS_MODEL?: string
	readonly FEATHERLESS_API_KEY?: string
	readonly FEATHERLESS_BASE_URL?: string
	readonly FEATHERLESS_MODEL?: string
	readonly PERPLEXITY_API_KEY?: string
	readonly PERPLEXITY_BASE_URL?: string
	readonly PERPLEXITY_MODEL?: string
	readonly MISTRAL_API_KEY?: string
	readonly MISTRAL_BASE_URL?: string
	readonly MISTRAL_MODEL?: string
	readonly MOONSHOT_API_KEY?: string
	readonly MOONSHOT_BASE_URL?: string
	readonly MOONSHOT_MODEL?: string
	readonly MODELSCOPE_API_KEY?: string
	readonly MODELSCOPE_BASE_URL?: string
	readonly MODELSCOPE_MODEL?: string
	readonly CLOUDFLARE_WORKERS_AI_MODEL?: string
	readonly OLLAMA_MODEL?: string
	readonly OLLAMA_EMBEDDING_MODEL?: string
	readonly LM_STUDIO_MODEL?: string
	readonly PLAYER2_MODEL?: string
	readonly PLAYER2_SPEECH_MODEL?: string
	readonly VLLM_MODEL?: string
	readonly VITE_AIRI_WS_URL?: string
	readonly DEFAULT_CHAT_PROVIDER?: string
	readonly DEFAULT_SPEECH_PROVIDER?: string
	readonly DEFAULT_TRANSCRIPTION_PROVIDER?: string

[nekomeowww]

You can include them.