Corrected model names (per Google AI docs):
- gemini-2.5-flash-image (was gemini-2.0-flash-preview-image-generation)
- gemini-3-pro-image-preview (Nano Banana Pro)
- gemini-3.1-flash-image-preview (Nano Banana 2)
UI: replaced text input with dropdown selector showing all 3 models
with human-readable descriptions.
Default changed to gemini-2.5-flash-image (fast, high-volume).
AI Assistant: generate_project_cover tool description updated to be
provider-agnostic (works with both DALL-E and Gemini).
Co-Authored-By: claude-flow <ruv@ruv.net>
Schema:
- SystemSettings: geminiApiKey, geminiModel, imageProvider fields
- imageProvider: "dalle" (default) or "gemini"
Gemini Client (packages/api/src/gemini-client.ts):
- Direct HTTP call to Gemini REST API with responseModalities: [TEXT, IMAGE]
- Returns base64 data URL
- Error parsing with user-friendly messages
Router (project.ts):
- generateCover: routes to DALL-E or Gemini based on imageProvider setting
- New isImageGenConfigured query returning { configured, provider }
Admin UI (SystemSettingsClient.tsx):
- "Image Generation" section with provider radio buttons (DALL-E / Gemini)
- Conditional fields: DALL-E config or Gemini API key + model
- Separate save button for image settings
Security:
- geminiApiKey sanitized in audit logs (SENSITIVE_FIELDS)
- API key stored server-side only, never sent to client
Co-Authored-By: claude-flow <ruv@ruv.net>
Hartmut