---
phase: 03-generate-node-refactor
plan: 02
type: execute
---

<objective>
Implement provider-specific image generation in API route.

Purpose: Route generation requests to correct provider (Gemini, Replicate, fal.ai) based on selected model.
Output: Working generate endpoint that dispatches to provider implementations.
</objective>

<execution_context>
~/.claude/get-shit-done/workflows/execute-phase.md
~/.claude/get-shit-done/templates/summary.md
</execution_context>

<context>
@.planning/PROJECT.md
@.planning/STATE.md
@.planning/phases/03-generate-node-refactor/03-01-SUMMARY.md
@src/app/api/generate/route.ts
@src/lib/providers/types.ts
@src/lib/providers/replicate.ts
@src/lib/providers/fal.ts

**Tech available:** ProviderInterface.generate(), provider registry
**Constraints:** Preserve existing Gemini logic, image output only
</context>

<tasks>

<task type="auto">
  <name>Task 1: Implement Replicate provider generate() method</name>
  <files>src/lib/providers/replicate.ts</files>
  <action>
  Implement generate() that currently throws "not implemented":
  1. POST to https://api.replicate.com/v1/predictions with { version: modelId, input: { prompt } }
  2. Poll GET /predictions/{id} until status is "succeeded" or "failed"
  3. Fetch output image URL and convert to base64 data URL
  4. Return GenerationOutput with type: "image"
  5. Skip image inputs for now (Phase 5 adds URL server)
  </action>
  <verify>TypeScript compiles</verify>
  <done>Replicate generate() calls prediction API and returns image</done>
</task>

<task type="auto">
  <name>Task 2: Implement fal.ai provider generate() method</name>
  <files>src/lib/providers/fal.ts</files>
  <action>
  Implement generate():
  1. POST to https://fal.run/{modelId} with { prompt, ...parameters }
  2. Use "Key {apiKey}" auth header
  3. Extract image URL from response, convert to base64
  4. Return GenerationOutput with type: "image"
  </action>
  <verify>TypeScript compiles</verify>
  <done>fal.ai generate() calls fal.run and returns image</done>
</task>

<task type="auto">
  <name>Task 3: Update generate API route for multi-provider dispatch</name>
  <files>src/app/api/generate/route.ts</files>
  <action>
  1. Accept both formats:
     - Legacy: { model, prompt, images, aspectRatio, resolution }
     - New: { selectedModel, prompt, images, parameters }
  2. If selectedModel provided and provider !== "gemini":
     - Get provider from registry
     - Get API key from request headers (X-Replicate-API-Key, X-Fal-API-Key)
     - Call provider.generate()
     - Return result
  3. Otherwise use existing Gemini logic
  4. Extract Gemini code into helper function for clarity
  </action>
  <verify>npm run build, test Gemini generation still works</verify>
  <done>API route dispatches to correct provider, Gemini unchanged</done>
</task>

</tasks>

<verification>
- [ ] `npm run build` succeeds
- [ ] Gemini generation works as before
- [ ] Replicate provider has generate() implementation
- [ ] fal.ai provider has generate() implementation
</verification>

<success_criteria>
- Generate API routes requests to correct provider
- Provider generate() methods implemented
- Existing Gemini functionality preserved
</success_criteria>

<output>
Create `.planning/phases/03-generate-node-refactor/03-02-SUMMARY.md`
</output>