Decision-Log.md
... ...
@@ -4,6 +4,7 @@ Use this page for decisions that change production canon. Newest entries go firs
4 4
5 5
| Date | Area | Decision | Status | Affected pages |
6 6
| --- | --- | --- | --- | --- |
7
+| 2026-07-26 | Initial production pipeline | Generate one approved Anima first frame per storyboard shot, animate it with local Wan 2.2 14B image-to-video, and generate music plus narration with Suno | Approved | [Visual Pipeline](/Visual-Pipeline), [Music and Voices](/Music-and-Voices), [Workflow Registry](/Workflow-Registry) |
7 8
| 2026-07-26 | Production bible | Repurpose the deprecated FaithOriginals.Media build wiki as The Lantern Club Production Bible | Approved | [Home](/) |
8 9
| 2026-07-26 | Series engine | Develop Little Lanterns as recurring alphabet seasons with Lumi and Pip | Candidate | [Little Lanterns](/Little-Lanterns), [Seasons](/Seasons) |
9 10
| 2026-07-26 | Episode format | Test a 4–5 minute Find → Meet → Help → Discover → Sing → Shine structure | Candidate | [Little Lanterns Format](/Little-Lanterns-Format) |
Episode-Template.md
... ...
@@ -40,14 +40,15 @@ Lumi and Pip meet **[guest]**, help with **[problem]**, and discover **[meaning]
40 40
41 41
**Two voice roles:**
42 42
**Spoken-word tags:**
43
+**Narration role:**
43 44
**Feeling arc:**
44 45
**Hook:**
45 46
**Suno record / link:**
46 47
47 48
## 12-panel storyboard
48 49
49
-| Panel | Shot | Story purpose | Prompt record | Seed | Review |
50
-| ---: | --- | --- | --- | ---: | --- |
50
+| Panel | Shot | Story purpose | Anima first frame | Wan motion | Review |
51
+| ---: | --- | --- | --- | --- | --- |
51 52
| 01 | Establishing wide | | | | |
52 53
| 02 | Medium | | | | |
53 54
| 03 | Insert / ECU | | | | |
... ...
@@ -63,10 +64,12 @@ Lumi and Pip meet **[guest]**, help with **[problem]**, and discover **[meaning]
63 64
64 65
## Generation provenance
65 66
66
-**Image model / checkpoint:**
67
-**Workflow version:**
67
+**Anima model / checkpoint:**
68
+**Anima workflow version:**
68 69
**Dimensions:** 864 × 1152 (3:4 portrait candidate)
69
-**Video model / workflow:**
70
+**Wan 2.2 14B workflow version:**
71
+**Suno music output:**
72
+**Suno narration output:**
70 73
**SmartGallery collection:**
71 74
**Approved asset IDs:**
72 75
Little-Lanterns-Format.md
... ...
@@ -3,6 +3,23 @@
3 3
**Status:** Candidate
4 4
**Target duration:** 4–5 minutes
5 5
6
+## Initial production format
7
+
8
+The episode is designed as one coordinated audio-visual sequence:
9
+
10
+1. **Anima storyboard** — generate and approve the first frame for every planned
11
+ shot.
12
+2. **Wan 2.2 14B image-to-video** — animate each approved first frame as a short
13
+ shot.
14
+3. **Suno music and narration** — generate the musical structure, two-character
15
+ voices, spoken narration, spoken-word tags, and emotional performance.
16
+4. **Edit** — assemble the Wan shots against the approved Suno audio and preserve
17
+ all prompt and workflow provenance.
18
+
19
+The 12-panel storyboard is therefore not disposable concept art. Each approved
20
+panel is the visual start frame and continuity contract for its corresponding
21
+video shot.
22
+
6 23
## Story rhythm
7 24
8 25
| Beat | Function | Typical visual language |
... ...
@@ -35,6 +52,10 @@ same composition. A standard 12-panel board should contain:
35 52
Every shot prompt must state framing, lens intent, camera height, subject scale,
36 53
foreground/background relationship, and who is visible.
37 54
55
+Each panel must also leave a clear motion opportunity for Wan: a character action,
56
+camera move, environmental movement, or emotional performance change. Avoid
57
+frames that are attractive but have no useful next action.
58
+
38 59
## Learning rule
39 60
40 61
Repeat the letter naturally in action, dialogue, and song. Do not turn the
Music-and-Voices.md
... ...
@@ -1,10 +1,14 @@
1 1
# Music and Voices
2 2
3
-**Status:** Candidate production standard
3
+**Status:** Approved initial tool direction; prompt format remains Candidate
4 4
5
-The working music pipeline is strongest when a song is designed for two distinct
6
-voices, uses explicit spoken-word tags, and describes the emotional movement of
7
-each section.
5
+Suno is the initial generation tool for both **music and narration**. The working
6
+pipeline is strongest when a piece is designed for two distinct voices, uses
7
+explicit spoken-word tags, and describes the emotional movement of each section.
8
+
9
+The audio concept is developed with the storyboard rather than added after
10
+animation. Its selected structure becomes the timing and emotional guide for the
11
+Wan image-to-video shots and final edit.
8 12
9 13
## Voice functions
10 14
... ...
@@ -39,11 +43,16 @@ song:
39 43
anchor_words: []
40 44
generation_record:
41 45
tool: Suno
46
+ output_mode: integrated_music_and_narration
42 47
prompt_version: ""
43 48
output_link: ""
44 49
status: Draft
45 50
```
46 51
52
+For episodes where control is stronger with separate generations, preserve
53
+distinct `music_output` and `narration_output` records and document how they are
54
+combined in the edit.
55
+
47 56
## Rules
48 57
49 58
* Give each voice an intention and feeling, not just a name.
... ...
@@ -51,3 +60,4 @@ song:
51 60
* Let the chorus express the discovery, not merely repeat the letter.
52 61
* Write visual beats and song structure together.
53 62
* Preserve generation prompts and selected output links.
63
+* Lock the chosen audio structure before final Wan shot timing and picture edit.
Visual-Pipeline.md
... ...
@@ -1,17 +1,72 @@
1 1
# Visual Pipeline
2 2
3
-**Status:** Candidate
3
+**Status:** Approved initial pipeline
4 4
5 5
## Production flow
6 6
7
-1. Lock the episode’s cast, guest, location, and style records.
7
+1. Lock the episode’s cast, guest, location, style, and audio intent.
8 8
2. Write a 12-panel board using the coverage rules in [Little Lanterns Format](/Little-Lanterns-Format).
9
-3. Compile each structured record through the model adapter.
10
-4. Generate contact-sheet candidates locally in ComfyUI.
11
-5. Review character recognition, location continuity, shot diversity, and story clarity.
12
-6. Approve image frames before image-to-video work.
13
-7. Generate short motion tests, then edit to music and voice.
14
-8. Store binaries in SmartGallery and record links and provenance here.
9
+3. Compile each structured image record through the Anima model adapter.
10
+4. Generate Anima storyboard candidates locally in ComfyUI.
11
+5. Review character recognition, location continuity, shot diversity, story
12
+ clarity, and the motion opportunity in every frame.
13
+6. Approve one Anima frame per shot. That image becomes the first frame and
14
+ visual continuity contract for image-to-video.
15
+7. Write a motion prompt that describes only what changes after the approved
16
+ first frame.
17
+8. Animate each approved frame through the local **Wan 2.2 14B
18
+ image-to-video** workflow.
19
+9. Generate the episode’s music, character voices, spoken narration,
20
+ spoken-word tags, and emotional performance through **Suno**.
21
+10. Edit the Wan shots to the approved Suno audio.
22
+11. Store binaries in SmartGallery and record links and provenance here.
23
+
24
+## Handoff contracts
25
+
26
+### Anima → Wan 2.2 14B
27
+
28
+Every approved storyboard shot hands off:
29
+
30
+```yaml
31
+shot_handoff:
32
+ shot_id: S01E__-SH__
33
+ approved_first_frame: ""
34
+ anima_prompt_record: ""
35
+ anima_workflow_version: ""
36
+ seed: ""
37
+ dimensions: 864x1152
38
+ visible_characters: []
39
+ continuity_anchors: []
40
+ intended_action: ""
41
+ camera_motion: ""
42
+ environmental_motion: ""
43
+ wan_motion_prompt: ""
44
+ wan_workflow_version: ""
45
+```
46
+
47
+The first-frame image owns appearance, framing, wardrobe, location, props, and
48
+starting pose. The Wan motion prompt owns change over time. It must not
49
+redesign the frame.
50
+
51
+### Suno → edit
52
+
53
+Every approved audio handoff records:
54
+
55
+```yaml
56
+audio_handoff:
57
+ episode: S01E__
58
+ suno_prompt_record: ""
59
+ music_output: ""
60
+ narration_output: ""
61
+ selected_version: ""
62
+ emotional_arc: ""
63
+ spoken_word_tags: []
64
+ edit_notes: ""
65
+```
66
+
67
+Music and narration may be generated as one integrated performance or as
68
+separate approved outputs, but the selected structure must be known before the
69
+final picture edit.
15 70
16 71
## Current image baseline
17 72
... ...
@@ -21,6 +76,9 @@
21 76
| Aspect ratio | 3:4 |
22 77
| Working dimensions | 864 × 1152 |
23 78
| Board size | 12 panels |
79
+| Storyboard / first-frame model | Anima |
80
+| Image-to-video model | Wan 2.2 14B |
81
+| Music and narration | Suno |
24 82
| Character strategy | Distinctive visible anchors + structured prompt inheritance |
25 83
| Review strategy | Contact sheet first, individual frames second |
26 84
... ...
@@ -34,6 +92,9 @@
34 92
* Location anchors recur without forcing identical backgrounds.
35 93
* Hands, gaze, props, and screen direction support the intended edit.
36 94
* No unexplained duplicate character or wardrobe mutation appears.
95
+* Every approved frame has a specific, achievable motion plan.
96
+* Wan preserves the approved first-frame identity, composition, and location.
97
+* Music and narration support the same emotional beats used by the storyboard.
37 98
38 99
## Known lesson
39 100
Workflow-Registry.md
... ...
@@ -6,9 +6,9 @@ review link whenever a workflow changes.
6 6
| Workflow | Purpose | Runtime | Version / path | Status |
7 7
| --- | --- | --- | --- | --- |
8 8
| Character reference pack | Turnaround, expressions, pair scale | Local ComfyUI | TBD | Draft |
9
-| 12-panel storyboard | Consistent cast, location, varied coverage | Local ComfyUI / Anima | Mary–Joseph V1 is the current learning reference | Candidate |
10
-| Image-to-video shot | Animate an approved keyframe | Local ComfyUI | TBD | Draft |
11
-| Music and voice | Two-voice song with spoken and feeling tags | Suno workflow | TBD | Candidate |
9
+| 12-panel first-frame storyboard | Consistent cast, location, varied coverage, and approved first frame for every shot | Local ComfyUI / Anima | Mary–Joseph V1 is the current learning reference | Approved direction |
10
+| Image-to-video shot | Animate an approved Anima first frame without redesigning it | Local ComfyUI / Wan 2.2 14B | Exact workflow path and version TBD | Approved direction |
11
+| Music and narration | Music, two-character voices, spoken narration, spoken-word tags, and feelings | Suno workflow | Exact prompt version TBD | Approved direction |
12 12
| Asset review | Contact sheets and approved output collections | SmartGallery | TBD | Candidate |
13 13
14 14
## Required workflow record