From a284757fdcd1de9d8bb1fdb26b2b5ecd9398b73e Mon Sep 17 00:00:00 2001
From: shrimbly <falloon.willie@gmail.com>
Date: Mon, 12 Jan 2026 21:04:00 +1300
Subject: [PATCH] docs(12): create phase plan

Phase 12: Model Improvements
- 1 plan created
- 2 total tasks defined
- Ready for execution
---
 .../12-model-improvements/12-01-PLAN.md       | 102 ++++++++++++++++++
 1 file changed, 102 insertions(+)
 create mode 100644 .planning/phases/12-model-improvements/12-01-PLAN.md

diff --git a/.planning/phases/12-model-improvements/12-01-PLAN.md b/.planning/phases/12-model-improvements/12-01-PLAN.md
new file mode 100644
index 00000000..06198d0d
--- /dev/null
+++ b/.planning/phases/12-model-improvements/12-01-PLAN.md
@@ -0,0 +1,102 @@
+---
+phase: 12-model-improvements
+plan: 01
+type: execute
+---
+
+<objective>
+Extend model cache TTL and verify Replicate image models work end-to-end.
+
+Purpose: Reduce API calls with longer cache duration and confirm Replicate image generation functions correctly.
+Output: Updated cache configuration and verified working Replicate image generation.
+</objective>
+
+<execution_context>
+~/.claude/get-shit-done/workflows/execute-phase.md
+~/.claude/get-shit-done/references/checkpoints.md
+</execution_context>
+
+<context>
+@.planning/PROJECT.md
+@.planning/ROADMAP.md
+@.planning/STATE.md
+
+# Key files:
+@src/lib/providers/cache.ts
+
+# Prior context:
+Model cache currently uses 10-minute TTL. Models don't change frequently, so longer TTL reduces API calls.
+Replicate image generation implemented in Phase 2-3, needs end-to-end verification.
+</context>
+
+<tasks>
+
+<task type="auto">
+  <name>Task 1: Extend model cache TTL to 1 hour</name>
+  <files>src/lib/providers/cache.ts</files>
+  <action>Change DEFAULT_TTL from 10 minutes (10 * 60 * 1000) to 1 hour (60 * 60 * 1000). Update the comment to reflect "1 hour" instead of "10 minutes". This reduces API calls to provider model listing endpoints.</action>
+  <verify>grep "DEFAULT_TTL" src/lib/providers/cache.ts shows 60 * 60 * 1000</verify>
+  <done>Cache TTL is 1 hour (3,600,000 ms)</done>
+</task>
+
+<task type="checkpoint:human-verify" gate="blocking">
+  <what-built>Extended cache TTL configuration</what-built>
+  <how-to-verify>
+    1. Start dev server: npm run dev
+    2. Open http://localhost:3000
+    3. Configure a Replicate API key in Settings if not already done
+    4. Click the Replicate icon in the floating action bar to open model search
+    5. Search for "flux" and select a text-to-image model (e.g., "black-forest-labs/flux-schnell")
+    6. Add a prompt like "a mountain landscape at sunset"
+    7. Click Run (Cmd+Enter) to generate an image
+    8. Confirm: Image appears in the output node
+    9. Confirm: No errors in console or UI
+  </how-to-verify>
+  <resume-signal>Type "approved" if Replicate image generation works, or describe any issues</resume-signal>
+</task>
+
+</tasks>
+
+<verification>
+Before declaring phase complete:
+- [ ] `npm run build` succeeds without errors
+- [ ] Cache TTL is set to 1 hour
+- [ ] Replicate image model generates successfully (human verified)
+</verification>
+
+<success_criteria>
+
+- Cache TTL extended to 1 hour
+- Replicate image generation verified working
+- Build passes
+- Phase 12 complete
+</success_criteria>
+
+<output>
+After completion, create `.planning/phases/12-model-improvements/12-01-SUMMARY.md`:
+
+# Phase 12-01: Model Improvements Summary
+
+**[Substantive one-liner - what shipped]**
+
+## Accomplishments
+
+- [Key outcome 1]
+- [Key outcome 2]
+
+## Files Created/Modified
+
+- `path/to/file.ts` - Description
+
+## Decisions Made
+
+[Key decisions and rationale, or "None"]
+
+## Issues Encountered
+
+[Problems and resolutions, or "None"]
+
+## Next Step
+
+Phase complete, ready for milestone v1.1 completion.
+</output>