photo-title
Skill4 starsChanged 10 days agoMITView on GitHub
Build an animated title whose letters are made of photos — like the "2026" in year-recap reels, where each stroke of each digit is a different photo snapping in one after another — matching a reference's exact letter shapes (traced pixel for pixel) or a font, with clean anti-aliased edges, rendered as frames and baked for FilmCraft. Use it whenever an intro or title should be "made of photos", "stacked from pictures", a photo collage in the shape of text or a year, or when the user says the title's font or build-up doesn't match the reference.
In this repository6 other files it gives its agents.
- analyze-referenceSkill
- check-renderSkill
- filmcraft-editingSkill
- match-gradeSkill
- recap-editSkill
- select-shotsSkill
The file
We split this skill into the rules it gives an agent
3 rules. About 0 are common, 2 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 46
Look at the finished frame and a few build-up frames side by side with the reference before going on.
- Distinctive: line 47
If a glyph's strokes split oddly, check the trace (specks) first.
- Very specific: line 51
Import with import_media(paths=[work/title/title_0000.png], imageSequence: true, bake: true) and put it at the start of V1 (plan intro).
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Photo-filled title
What users notice: the letter shape (it must look like the reference's font, not a similar one), the build-up (pieces snapping in stroke by stroke in the reference's order and timing — a static title "with stacking sound" is a failure), clean edges ("rough edges" read as unfinished), and photos staying inside their stroke (no overflow across seams). Scripts need Pillow.
1. Study the reference title
With analyze-reference, extract the intro at full scale and look at:
- the finished title frame (shape, size as a fraction of the frame height, spacing)
- frames every 2–3 frames through the build-up: which piece appears when (pen order), how many frames between pieces, whether the first frame flashes the finished title, when each glyph starts
- whether photos inside are upright or sideways
2. Get the exact letter shape
Best: trace it from the reference's finished frame (lettering must sit on a dark background):
python3 ${CLAUDE_SKILL_DIR}/scripts/trace_mask.py ref_title_frame.png work/title_mask.png --box x0,y0,x1,y1--box must contain only the lettering (no caption). Dark areas of the reference's own photos become
specks; small ones are filled automatically (raise --speck if holes remain, but not above the size of
real counters).
Otherwise use a font (--font /path/Font.ttf --text 2026); pick a heavy condensed face close to the
reference.
3. Render the animated title
python3 ${CLAUDE_SKILL_DIR}/scripts/photo_title.py --photos <photos dir> --out work/title \
--mask work/title_mask.png --sideways --glyph-starts 6,22,40,54 --step 3 --frames 105 \
--first-frame-full --reuse 2=0- outline cleaned at 2x and anti-aliased; glyphs split into strokes along their own edges (bars across,
strokes down), one photo per stroke, revealed top to bottom every
--stepframes from each glyph's start --reuse 2=0: a repeated character (the second "2") takes the cleaner trace of the first--sideways: photos rotated 90° as many reels do; omit for upright--height-fracletter height vs frame (match the reference);--width/--heightyour sequence size Look at the finished frame and a few build-up frames side by side with the reference before going on. If a glyph's strokes split oddly, check the trace (specks) first.
4. Use it
Import with import_media(paths=[work/title/title_0000.png], imageSequence: true, bake: true) and put
it at the start of V1 (plan intro). Never re-write frames of an already-imported sequence in place —
the app can keep showing stale frames; write a new folder and bake again. Confirm in the final render
with check-render --intro-frames N that the intro actually animates.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

