# Model Split — Implementation Plan

> **For agentic workers:** REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development (recommended) or superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox (`- [ ]`) syntax for tracking.

**Goal:** Rename ai_model_writing → ai_model_websearch, split provider models into scraping and websearch lists with GPT-5 generation models.

**Architecture:** Rename DB column, split JSONB `models` into `models_scraping` + `models_websearch`, update all backend structs/queries/handlers + frontend types/dropdowns.

**Tech Stack:** Rust (sqlx), SolidJS, PostgreSQL

**Spec:** `docs/superpowers/specs/2026-03-25-model-split-design.md`

---

### Task 1: Migration — rename column + restructure JSONB

**Files:**
- Create: `backend/migrations/20260325000019_model_split.sql`
- Modify: `CLAUDE.md`

The migration:
1. Renames `ai_model_writing` → `ai_model_websearch` in settings
2. Renames `models` → `models_scraping` in admin_providers JSONB
3. Copies `models_scraping` → `models_websearch` (same models initially for Gemini/Anthropic)
4. Updates OpenAI specifically with GPT-5 model lists

### Task 2: Backend model + DB changes

Update all Rust structs and DB queries for the rename and new JSONB structure.

### Task 3: Pipeline + handlers

Update synthesis.rs, config handler, admin handler.

### Task 4: Frontend

Update types, Settings dropdowns, i18n, E2E test.

---

These 4 tasks are tightly coupled — I'll execute them as 2 subagent calls (backend + frontend).