Motion Creative Strategy Bootcamp — Corpus

13K marketers just finished our 7-week creative strategy bootcamp. Now your agent can take it too.

The complete bootcamp — every Tuesday class, every Thursday coaching call — extracted into structured markdown for any agent that reads files. Drop into Claude, Cursor, Runneth, or whatever you're using.

Scope: 13 sessions × 2 files each (structured lesson + verbatim reference). ~80K tokens of lessons + 25–100K tokens per reference file. Every framework, claim, tactic, anti-pattern, key quote, ad shown, and slide — extracted, indexed, and cross-referenced.

Shape

bootcamp/
├── index.md              ← cheat sheet — load this first
├── week-01/
│   ├── tuesday-evan-what-is-creative-strategy.md       ← lesson (~5–7K tokens)
│   ├── tuesday-evan-what-is-creative-strategy.full.md  ← reference (~25–100K tokens)
│   ├── thursday-coaching-hook-writing.md
│   └── thursday-coaching-hook-writing.full.md
├── week-02/  …  week-07/

Every session has two files:

File	Contains	Open when
<session>.md (lesson)	Frontmatter, summary, chapters, frameworks with full taxonomies, claims, tactics, anti-patterns, named entities (people / brands / tools), key quotes, time-bound claims, homework, cross-week refs. Coaching sessions also include student questions answered, ads critiqued live, live worked examples.	By default. Answers ~80% of questions about a session.
<session>.full.md (reference)	Frontmatter (copy) + verbatim ad-facts registry (every ad shown, with brand / format / hook / on-screen text / speaker framing) + verbatim slide-facts registry (every distinct slide with title / layout / body / charts / annotations / reveal states / speaker framing) + full speaker-tagged transcript with inline [VISUAL: …] annotations.	When you need a verbatim quote, an exact slide, or to cite a specific moment by timestamp.

Total corpus: ~330 KB (~80K tokens) for the lesson layer; per-session reference files run 25–100K tokens each.

Quick start

Claude (web/desktop) — use the GitHub connector and point it at Motion-Creative/bootcamp. Or drop raw file URLs into the chat:

https://raw.githubusercontent.com/Motion-Creative/bootcamp/main/index.md

Claude Code / Cursor / any local agent:

git clone https://github.com/Motion-Creative/bootcamp.git

Add it to your project's context. Load index.md first — it's the router.

Runneth or any markdown-aware agent: clone, mount, point it at the folder. It's just markdown.

Speakers

Sarah Levinger (Tether Insights), Dara Denney (Point Guard Media), Will Sartorius (SelfMade), Jade Heritage (Calm), Daniel Rivera (Harry's), Sophia Beauvoir (Ready Set), Viti Videtta (Happy Mammoth), Janae LeVander (Caraway), Jake Abrams (Odyssey), Joanna Wallace (Creativly), Eric Philippou (Laughing Heart Media), Sarah Burnett, Matteo Friend, Darcy Tennant (Heights), Blaise Ormond (MTRX Media), Naomi Peh Haeger (Spacegoods), and the Motion team.

Coverage

13 sessions across weeks 1–7 of the 8-week curriculum:

• 7 Tuesday classes (single-instructor or guest-panel deep dives, slide-driven)
• 6 Thursday group coaching calls (live student work review + tactic walkthroughs)

Not yet captured (sessions either un-recorded at extraction time or future-dated): Week 7 Thursday, Week 8 Tuesday (Sprint #3 capstone), Master's Series Weeks 9–10.

How an agent should use this

1. Always load index.md first. It contains:

   • A 1-paragraph overview per week
   • A card per session (instructor, frameworks introduced, distinctive content, "what to open this file for")
   • A topic index (hooks → which sessions, Andromeda → which sessions, etc.)
   • A speaker index (every instructor, guest, and recurring coach)
   • A framework index (every named framework → canonical home session)
   • A "If a user asks…" routing table (~35 likely Slack questions → which file to open first)

2. Open the lesson .md for the relevant session. That handles most questions.

3. Only open .full.md when you need verbatim wording, exact slide content, or to cite a specific moment. Reference files are large; don't load them speculatively.

4. Cross-reference — many sessions reference each other (e.g., Week 7 Tuesday repeatedly cites Dara's Week 3 frameworks). The ## Cross-week references section in each lesson file lists these explicitly.

Notable known gaps

• Week 5 Thursday's structured slide registry is a placeholder. The transcript captures slide content via 112 inline [VISUAL: …] annotations, but the dedicated slide registry pass exhausted retries. Slide content is recoverable via the transcript; only the deduplicated structured registry format is missing.
• Verification notes appear at the end of each lesson file documenting any corrections made during synthesis (name spellings, transcript loop artifacts, OCR quirks).

Provenance

• Videos: original Drive recordings, pre-encoded to 720p / 1.5 Mbps for upload.
• Extraction stack: Gemini 2.5 Pro for 4 parallel passes (verbatim transcript / slide registry / ad registry / structured metadata) at MEDIA_RESOLUTION_LOW to fit ~2.5h videos under the 1M-token input cap; Claude Opus 4.7 for verification + metadata synthesis using sampled key frames.

License

No formal license. Bootcamp recordings, slides, and frameworks belong to their respective speakers (Motion team + guest instructors). Transcripts and structured extraction are published here for fair-use reading and personal/agent consumption — drop them into your tools, learn from them, share what you've learned. For redistribution or commercial use of speaker content, please reach out.