About Book Mascots

This site collects pedagogical mascots used across a family of intelligent textbooks. Each mascot is a small cartoon character with a fixed visual identity and a set of standard "poses" — neutral, welcome, thinking, tip, encouraging, warning, celebration — that authors drop into chapters to add personality and to give students consistent visual cues.

Why pedagogical mascots matter

A mascot is not decoration. When used well, it does real instructional work.

• Reduced cognitive load. A recurring character is a known visual landmark. When a student sees Olli the Octopus mid-chapter, they don't have to spend attention parsing a new illustration — they know who this is, and they know what kind of moment it signals.
• Tone shift without prose. A "warning" pose communicates be careful here faster and more gently than a paragraph of italicized prose. A "celebration" pose at the end of a chapter rewards completion in a way text alone can't.
• Emotional pacing. Long-form technical material is grueling. A friendly face every few pages keeps the reader's affect warm and signals that the author is on the student's side.
• Identity for the book. Sparkle the Unicorn, Bailey the Beaver, Axiom the Owl — these become the face of a textbook. Students remember the mascot long after they've forgotten a specific definition, and the mascot becomes the on-ramp back into the material.
• Inclusivity and approachability. A cartoon animal carries no gender, age, or cultural baggage in the way a human illustration does. It lowers the barrier for readers who don't see themselves in stock photography of "the typical student."
• Consistency across surfaces. The same mascot poses can appear in the textbook, the lecture slides, the quiz feedback, the chatbot avatar, and the social-media announcement. That continuity reinforces the book as a coherent product rather than a pile of files.

The poses themselves are a small visual vocabulary. Once a reader has learned that thinking means "pause and reflect" and tip means "here is a shortcut," every subsequent chapter inherits that vocabulary for free.

How a book mascot is created

The workflow we use across these books is roughly:

1. Pick a character concept. Match the species, color palette, and props to the subject matter. Bailey the Beaver for ecology (a builder of habitats). Sentinel the Fox for cybersecurity ("trust, but verify"). Delta — a triangle robot — for calculus. The link between mascot and subject is part of what makes it memorable.
2. Write a Character Sheet. A short prose document — name, species, color hex codes, props, personality, "what the hands are doing," background rules. This is the source of truth that every pose prompt re-anchors to. Without it, drift across poses is guaranteed.
3. Author seven pose prompts. Each pose prompt embeds the full character sheet, then adds two or three sentences describing the pose-specific gesture (waving for welcome, hand-on-chin for thinking, etc.). Self-contained prompts matter — image generators do not reliably maintain character identity across a session.
4. Generate, trim, save. Run each prompt through an image generator (DALL-E, Midjourney, Imagen, Firefly), trim padding, save as a transparent PNG.
5. Wire into the book. Copy the PNGs into docs/img/mascot/ and reference them from chapter markdown using a small shortcode or admonition pattern.

The expensive step is #3 and #4. Getting seven poses that look like the same character — same proportions, same eye color, same prop placement — typically takes many regenerations and several rounds of prompt tightening. This is where most of the cost lives.

Token costs and image-generation costs

A complete mascot set has real cost behind it, and the cost is dominated by image generation rather than text tokens.

• Character-bible drafting: a few thousand input/output tokens. Trivial.
• Pose prompt authoring: seven self-contained prompts, each repeating the bible. Maybe 5–10K tokens of LLM work total.
• Image generation: seven poses × 3–10 regenerations each to lock in consistency. At commercial rates (a few cents to ~25¢ per image depending on the tool and tier), a finished mascot easily runs $5–$30 in image-API spend, plus the human time to triage the bad outputs.
• Touch-up: background-trimming, centering, occasional manual edits in an image editor.

Across this site that is 20+ mascots × seven poses each = 140+ generated images. The aggregate cost — in money, in API budget, and especially in iteration time — is non-trivial.

Not every book needs a unique mascot

A counterintuitive lesson from running this experiment across a couple dozen books: most books do not need their own mascot. The marginal pedagogical value of "a brand-new character for this book" is usually small once a good general-purpose mascot exists. What students actually benefit from is the consistency of the pose vocabulary — neutral, welcome, thinking, tip — not the novelty of the species.

Good defaults:

• A course series (e.g. all your math books, or all your science books) can share one mascot. Continuity across volumes is a feature, not a bug.
• A book with a strong subject-specific identity — moss biology, ecology, cybersecurity — is a good candidate for a custom mascot, because the species/prop choice carries real meaning.
• A book in a generic domain — "Introduction to Programming," "Study Skills," a tutorial — is usually well served by an existing general-purpose mascot.

Treat custom mascots as a deliberate investment, not a default deliverable.

Reuse: Claude Code, no image-LLM license required

This is the practical payoff of publishing the mascots here. Once a mascot exists in this repo, anyone with a Claude Code license can reuse it without paying for a text-to-image model.

Concretely, an author working in Claude Code can:

• Pull the seven PNGs for Olli the Octopus (or any mascot here) directly into their book's docs/img/mascot/ directory.
• Wire them into chapter markdown the same way the source books do.
• Skip the entire image-generation pipeline — no Midjourney subscription, no OpenAI Images API key, no Firefly seat, no waiting on regenerations.

This matters for two reasons:

1. Lower the barrier to making a textbook feel finished. A first-time author working on a small intelligent textbook should not have to set up a second AI subscription just to get a friendly face into their chapters.
2. Standardize the pose vocabulary. When multiple books pull from the same mascot library, students who read more than one of them benefit from cross-book consistency in what thinking and warning and celebration look like.

The mascots in this repo are licensed for reuse across our intelligent-textbook projects. Pick one, copy the seven PNGs, and ship.

What this site is

• A gallery (Mascot Gallery) of every mascot, showing the neutral pose and the character name.
• A detail page per mascot showing all seven poses side-by-side.
• The original prompt files (image-prompts.md) under each mascot directory, so you can regenerate, fork, or adapt a mascot for a related book.
• An update-mascots skill that scans sibling book projects and brings new mascots into this gallery automatically.

The goal is a small, reusable, low-friction set of pedagogical characters that any author in our orbit can lift wholesale and put to work.