chore: release 1.2.3

calliope-backend/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "calliope"
-version = "1.2.1"
+version = "1.2.3"
 description = "Local-first AI story-to-video studio"
 requires-python = ">=3.11"
 dependencies = [

calliope-web/package.json

@@ -1,6 +1,6 @@
 {
-  "name": "calliope-web",
-  "version": "1.2.1",
+	"name": "calliope-web",
+	"version": "1.2.3",
   "private": true,
   "scripts": {
     "dev": "vite dev",

docs/changelogs/2026-08-29.md

@@ -1,6 +1,6 @@
 # Calliope changelog — 29 August 2026
 
-**Per-agent LLMs + clip-native export** update. Development branch `test-pr-25`.
+**Per-agent LLMs + clip-native export + video-extend continue** update. Released as **v1.2.3**. Development branch `test-pr-25`.
 
 Full changelog index: [README](./README.md). Previous: [2026-08-24](./2026-08-24.md).
 
@@ -54,6 +54,7 @@ Added:
 - **Video stage:** continue scenes require a workflow with an `(Input:video)` input — otherwise Generate is disabled with a warning. When valid, a clip source picker offers **Auto** (previous clip, default), **Upload file**, or **From timeline** (explicit earlier clip).
 - **Enqueue:** resolution order is explicit value → nearest earlier scene's clip → `continue_source` payload for job-time resolution.
 - **Worker:** `_resolve_continue_source` picks up the nearest earlier scene's clip at run time (safe for batch enqueues — the queue is concurrency-1), writes it into the `(Input:video)` node, and uploads it to ComfyUI. No ffmpeg involved.
+- **Clip source modal:** the source picker is a full modal, not a cramped dropdown — **Auto (previous clip)** and **Upload file** cards plus a thumbnail grid of **any other scene's clip in the project** (not just earlier ones, not just assets). Picking a clip fills the workflow's video input; Auto defers resolution to the queue.
 
 Migration:
 
@@ -61,6 +62,10 @@ Migration:
 - Workflows relying on `clipindex`-titled Motion Context nodes lose that auto-fill; re-tag the clip input with `(Input:video)`.
 - For the H3 extend pattern, target [kat3ri/ComfyUI-MiniMax-H3-Extend](https://github.com/kat3ri/ComfyUI-MiniMax-H3-Extend): `LoadVideo (Input:video)` → MiniMax H3 encode/extend patched nodes → `(Output:video)`.
 
+## Fixed
+
+- **"Choose Video" modal could render the wrong tab content** (e.g. the header showing Clips/Uploads while the body said "No character sheets yet"). Two causes fixed: the same file appearing twice in the picker grid (a scene clip that was also uploaded) crashed Svelte's keyed rendering mid-draw — options are now deduped by path; and a tile whose media kind changes could keep a stale active tab — the modal now clamps the tab to the visible list.
+
 ## Housekeeping
 
 - Added `.gitignore` files inside `calliope-backend/` and `calliope-web/` (venv, caches, SQLite data, node_modules, build output) so each sub-folder is self-protecting in addition to the repo-root rules.
@@ -71,4 +76,6 @@ Migration:
 - New config knob: `agent_llm_assignments` (role → profile id or null), exposed via `GET/POST /api/settings`.
 - `POST /api/settings` accepts `agent_llm_assignments` with server-side validation.
 - Export: `probe()` now returns `fps` (from ffprobe `r_frame_rate`), `parse_rate()` parses rational rates like `24000/1001`, `target_fps()` votes the conform rate, and `build_ffmpeg_cmd(..., fps=...)` takes an optional override.
+- New `ClipSourceModal.svelte` component backs the video source picker; `Modal.svelte` gained an `xl` size (1180px).
+- Version bumped to **1.2.3** (`calliope-web/package.json`, `calliope-backend/pyproject.toml`).
 - Full backend suite: **248 passed**.

docs/changelogs/README.md

@@ -2,6 +2,6 @@
 
 Dated notes for GitHub releases. Newest first.
 
-- [2026-08-29](./2026-08-29.md) — Model per agent + clip-native export fps
+- [2026-08-29](./2026-08-29.md) — Model per agent + clip-native export fps + video-extend continue (1.2.3)
 - [2026-08-24](./2026-08-24.md) — Agent + Assets Management (1.2.1)
 - [2026-08-23](./2026-08-23.md) — 1.2.0 / 1.2.1 video, H3, picker, release notes