feat: extend /api/generate-classroom to support full generation features (web search, image, video, TTS)

[cosarah]

Background

The current /api/generate-classroom API only accepts 3 parameters:

interface GenerateClassroomInput {
  requirement: string;
  pdfContent?: { text: string; images: string[] };
  language?: string;
}

However, the web UI's client-side generation pipeline already supports additional features:

• Web search — enriches outline generation with real-time web context (UserRequirements.webSearch)
• Image generation — generates images for slides via imageGenerationEnabled + image providers
• Video generation — generates video assets via videoGenerationEnabled + video providers
• TTS generation — generates narration audio via TTS providers

These features are passed via custom headers (x-image-generation-enabled, x-video-generation-enabled, etc.) in the client-side flow but are completely missing from the server-side generateClassroom() pipeline used by the API.

Goal

Extend GenerateClassroomInput to support these feature toggles so that API callers (e.g., OpenClaw) can control them:

interface GenerateClassroomInput {
  requirement: string;
  pdfContent?: { text: string; images: string[] };
  language?: string;
  // New fields
  enableWebSearch?: boolean;    // Enable web search for richer context
  enableImageGeneration?: boolean; // Generate images for slides
  enableVideoGeneration?: boolean; // Generate video assets
  enableTTS?: boolean;           // Generate narration audio (TTS)
}

Implementation Tasks

1. Extend GenerateClassroomInput interface

• Add the 4 optional boolean fields to lib/server/classroom-generation.ts

2. Wire up features in generateClassroom()

• Web search: Pass webSearch flag to UserRequirements → used by generateSceneOutlinesFromRequirements()
• Image generation: Pass imageGenerationEnabled to outline generator options, integrate media orchestrator for image generation after scenes are built
• Video generation: Pass videoGenerationEnabled to outline generator options, integrate media orchestrator for video generation
• TTS: Add TTS generation step after scene content is finalized (similar to client-side generateTTSForScene())

3. Update API route

• Accept new fields in app/api/generate-classroom/route.ts
• Pass them through to generateClassroom()

4. Feature availability detection

• Server already auto-detects available providers via getServerImageProviders(), getServerVideoProviders(), getServerTTSProviders(), getServerWebSearchProviders() in lib/server/provider-config.ts
• If a feature is requested but no provider is configured, either gracefully skip or return an error

5. Update OpenClaw skill documentation

skills/openmaic/references/generate-flow.md

• Document new optional fields in the request body
• Add examples showing how to enable each feature

skills/openmaic/references/provider-keys.md

• Add guidance for configuring image, video, TTS, and web search providers
• Explain which env vars to set for each capability

skills/openmaic/SKILL.md

• Local mode: During setup (Phase 3), ask the user which optional features they want to enable (web search, image gen, video gen, TTS), then guide them to configure the corresponding provider keys
• Hosted mode: Use the live demo's default configuration (no need to ask — features are pre-configured server-side)

6. Add capabilities endpoint (optional)

Consider adding a GET /api/capabilities or extending /api/health to return which features are available based on current server config:

{
  "status": "ok",
  "capabilities": {
    "webSearch": true,
    "imageGeneration": false,
    "videoGeneration": false,
    "tts": true
  }
}

This would help OpenClaw auto-detect what's available without trial and error.

Key Files

File	Role
lib/server/classroom-generation.ts	GenerateClassroomInput interface + generateClassroom() pipeline
app/api/generate-classroom/route.ts	API route handler
lib/types/generation.ts	UserRequirements type (already has webSearch field)
lib/generation/outline-generator.ts	Accepts imageGenerationEnabled / videoGenerationEnabled options
lib/server/provider-config.ts	Server-side feature availability detection
lib/hooks/use-scene-generator.ts	Client-side reference impl for all features
skills/openmaic/SKILL.md	OpenClaw skill SOP
skills/openmaic/references/generate-flow.md	Generation API documentation
skills/openmaic/references/provider-keys.md	Provider configuration guide

[viankie090912-commits]

使用OpenClaw生成的互动课堂，Hosted mode , 没有旁白声音。