docs(03-02): complete provider-specific execution plan

Tasks completed: 3/3
- Implement Replicate provider generate() method
- Implement fal.ai provider generate() method
- Update generate API route for multi-provider dispatch

SUMMARY: .planning/phases/03-generate-node-refactor/03-02-SUMMARY.md

.planning/ROADMAP.md

@@ -53,7 +53,7 @@ Plans:
 
 Plans:
 - [x] 03-01: Rename NanoBanana to GenerateImage, add model selector (image models only)
-- [ ] 03-02: Provider-specific execution in generate API route
+- [x] 03-02: Provider-specific execution in generate API route
 - [ ] 03-03: Backward compatibility for existing workflows
 
 ### Phase 4: Model Search Dialog
@@ -98,7 +98,7 @@ Phases execute in numeric order: 1 → 2 → 3 → 4 → 5 → 6
 |-------|----------------|--------|-----------|
 | 1. Provider Infrastructure | 2/2 | Complete | 2026-01-09 |
 | 2. Model Discovery | 3/3 | Complete | 2026-01-09 |
-| 3. Generate Node Refactor | 1/3 | In progress | - |
+| 3. Generate Node Refactor | 2/3 | In progress | - |
 | 4. Model Search Dialog | 0/2 | Not started | - |
 | 5. Image URL Server | 0/2 | Not started | - |
 | 6. Video & Polish | 0/4 | Not started | - |

.planning/STATE.md

@@ -10,18 +10,18 @@ See: .planning/PROJECT.md (updated 2026-01-09)
 ## Current Position
 
 Phase: 3 of 6 (Generate Node Refactor)
-Plan: 1 of 3 in current phase
+Plan: 2 of 3 in current phase
 Status: In progress
-Last activity: 2026-01-09 - Completed 03-01-PLAN.md (Rename NanoBanana to GenerateImage)
+Last activity: 2026-01-09 - Completed 03-02-PLAN.md (Provider-specific execution)
 
-Progress: [======================================] 40% (6/15 plans)
+Progress: [==========================================] 47% (7/15 plans)
 
 ## Performance Metrics
 
 **Velocity:**
-- Total plans completed: 6
-- Average duration: 5.5 min
-- Total execution time: 0.55 hours
+- Total plans completed: 7
+- Average duration: 5.3 min
+- Total execution time: 0.62 hours
 
 **By Phase:**
 
@@ -29,10 +29,10 @@ Progress: [======================================] 40% (6/15 plans)
 |-------|-------|-------|----------|
 | 1. Provider Infrastructure | 2/2 | 14 min | 7 min |
 | 2. Model Discovery | 3/3 | 14 min | 4.7 min |
-| 3. Generate Node Refactor | 1/3 | 5 min | 5 min |
+| 3. Generate Node Refactor | 2/3 | 10 min | 5 min |
 
 **Recent Trend:**
-- Last 5 plans: 4 min, 5 min, 5 min, 5 min
+- Last 5 plans: 5 min, 5 min, 5 min, 5 min
 - Trend: stable
 
 ## Accumulated Context
@@ -57,6 +57,9 @@ Recent decisions affecting current work:
 - Provider dropdown shows Gemini always, others only if API key configured
 - Aspect ratio/resolution controls shown only for Gemini provider
 - Backward compatibility via aliases: NanoBananaNode, saveNanoBananaDefaults
+- Server-side provider execution in API route (not client-side)
+- Header-based API key passing: X-Replicate-API-Key, X-Fal-API-Key
+- fal.ai sync API (fal.run) instead of queue-based async
 
 ### Deferred Issues
 
@@ -69,6 +72,6 @@ None.
 ## Session Continuity
 
 Last session: 2026-01-09
-Stopped at: Completed 03-01-PLAN.md (GenerateImageNode with provider/model selector)
+Stopped at: Completed 03-02-PLAN.md (Provider-specific execution)
 Resume file: None
-Next action: Execute 03-02-PLAN.md (Provider-specific execution in generate API route)
+Next action: Execute 03-03-PLAN.md (Backward compatibility for existing workflows)

.planning/phases/03-generate-node-refactor/03-02-SUMMARY.md

@@ -0,0 +1,106 @@
+---
+phase: 03-generate-node-refactor
+plan: 02
+subsystem: api
+tags: [replicate, fal-ai, image-generation, multi-provider]
+
+# Dependency graph
+requires:
+  - phase: 03-01
+    provides: GenerateImageNode with provider/model selector, SelectedModel type
+provides:
+  - Replicate provider generate() implementation with polling
+  - fal.ai provider generate() implementation
+  - Multi-provider dispatch in /api/generate route
+affects: [04-model-search-dialog, 05-image-url-server]
+
+# Tech tracking
+tech-stack:
+  added: []
+  patterns:
+    - Provider dispatch pattern in API route via provider-specific helpers
+    - Header-based API key passing (X-Replicate-API-Key, X-Fal-API-Key)
+    - Replicate prediction polling pattern
+
+key-files:
+  created: []
+  modified:
+    - src/lib/providers/replicate.ts
+    - src/lib/providers/fal.ts
+    - src/app/api/generate/route.ts
+
+key-decisions:
+  - "Server-side provider execution (helpers in route.ts) rather than client-side"
+  - "Header-based API key passing to avoid localStorage on server"
+  - "fal.ai sync API (fal.run) instead of queue-based async"
+
+patterns-established:
+  - "Provider-specific helper functions: generateWithGemini(), generateWithReplicate(), generateWithFal()"
+  - "Replicate polling with 1s interval until succeeded/failed"
+  - "Multi-format fal.ai response handling (images array, image object, output string)"
+
+issues-created: []
+
+# Metrics
+duration: 5 min
+completed: 2026-01-09
+---
+
+# Phase 3 Plan 2: Provider-Specific Execution Summary
+
+**Multi-provider dispatch in /api/generate with Replicate prediction polling and fal.ai sync API**
+
+## Performance
+
+- **Duration:** 5 min
+- **Started:** 2026-01-09T07:29:01Z
+- **Completed:** 2026-01-09T07:34:10Z
+- **Tasks:** 3
+- **Files modified:** 3
+
+## Accomplishments
+
+- Replicate generate() implementation with version fetching and prediction polling
+- fal.ai generate() implementation with sync fal.run API and multi-format response handling
+- Generate API route now dispatches to correct provider based on selectedModel.provider
+- Gemini logic extracted into generateWithGemini() helper for clarity
+
+## Task Commits
+
+Each task was committed atomically:
+
+1. **Task 1: Implement Replicate provider generate()** - `cbfdc9f` (feat)
+2. **Task 2: Implement fal.ai provider generate()** - `d9c0ce7` (feat)
+3. **Task 3: Update generate API route for multi-provider dispatch** - `ac11c6e` (feat)
+
+**Plan metadata:** (this commit)
+
+## Files Created/Modified
+
+- `src/lib/providers/replicate.ts` - Added ReplicatePrediction interface and generate() with API polling
+- `src/lib/providers/fal.ts` - Added FalGenerationResponse interface and generate() with sync API
+- `src/app/api/generate/route.ts` - Added multi-provider dispatch and provider-specific helper functions
+
+## Decisions Made
+
+- Server-side provider execution in API route rather than client-side provider.generate() calls
+- API keys passed via request headers (X-Replicate-API-Key, X-Fal-API-Key) to avoid localStorage access on server
+- fal.ai uses sync fal.run API for simplicity (queue-based async not needed for image generation)
+
+## Deviations from Plan
+
+None - plan executed exactly as written.
+
+## Issues Encountered
+
+None
+
+## Next Phase Readiness
+
+- Provider generation working for Replicate and fal.ai
+- Image inputs not yet passed to providers (requires Phase 5 URL server)
+- Ready for 03-03-PLAN.md (backward compatibility for existing workflows)
+
+---
+*Phase: 03-generate-node-refactor*
+*Completed: 2026-01-09*