How EduGenius's Content-Format Picker Caps What You Can Combine in One Generation
EduGenius's format picker — the grid you see before you generate anything — isn't a simple multi-select. It runs a live load-budget calculation against every format you tap, hard-caps combined selections at three, forces certain formats to generate alone, and gates a format's own configuration options (like MCQ difficulty) independently of that budget. A teacher can combine a worksheet and a study guide freely, but adding a long-format exam forces the request back down to one format, generated alone.
Quick Answer:
components/OutputFormatSelector.jsshows 15 formats to learners and 16 to teachers (one extra, teacher-only format) out of a 29-entry registry.lib/generation-complexity-policy.jsscores every selected format's complexity, credit cost, duration, and quantity into a load score, capping combined selections at three (MAX_GENERATION_SELECTIONS) and forcing some formats — Long Format Exam always, Presentation Slides above 12 slides, Image Creations above 3 images, Data Visualization above 5 charts — to generate alone. Credit cost is a separate gate layered on top. And the picker's own teacher hero-format list names six formats, but two of them (Rubric Grid, Presentation Slides) are flaggedhidden: truein the format registry, so only four actually render.
This piece is about the screen you see before anything exists — the one that decides what gets generated, not the dropdown that decides what file format it exports to afterward. Confusing the two is an easy mistake, since both get called "format" in casual conversation. What follows covers what the picker actually is, the load-budget system that's the real engineering behind it, the separate MCQ and credit gates, and the honest gap in what a teacher actually sees on screen.
What the Format Picker Actually Is
The format picker is the card grid at the top of /generate, rendered by OutputFormatSelector.js and mounted directly inside ContentGenerator.js. A teacher or learner taps one or more format cards — Quiz, Flashcards, Worksheet, Study Guide, and so on — before typing or uploading a single word of source material. Nothing gets created until that selection and the source content both exist.
The component's own code comment calls this grid "the product x-factor": primary formats sit flat and visible, everything else is one "+more" toggle away, and a format's configuration controls (question count, difficulty, slide count) unfold inline the moment you tap it — no second click, no separate settings screen (
components/ContentGenerator.js:859-863).
The Real Format Count: 29 in the Registry, 15 or 16 on Screen
The full catalogue lives in config/content-formats.js as a single CONTENT_FORMATS object. Counting every key — from multiple_choice_quiz down to anecdote — comes to 29 total format entries. Most of those never reach a teacher or learner at all.
Two filters decide what actually renders in the grid: a format's tier (1 or 2 to be generation-eligible by default) and whether it's flagged hidden: true. A third category, TEACHER_ONLY_FORMAT_IDS, adds specific teaching artefacts back in for teachers regardless of tier — but only if they aren't also hidden.
| Persona | Formats shown | Source |
|---|---|---|
| Learner | 15 | getLearnerGenerateFormatEntries() — tier ≤ 2, not hidden, not teacher-only |
| Teacher | 16 | The same 15, plus Leveled Homework (the one teacher-only format that isn't hidden) |
That "15+ content formats" line matches marketing copy elsewhere on this site precisely — a rarer thing to confirm cleanly than you'd expect in a corpus that's caught a few looser claims before. Everything else in the registry — nine formats including Rubric Grid, Newsletter, Data Visualization, and the pricier Slide Deck format — is either hidden, teacher-only-and-hidden, or excluded by tier, and none of it appears in the grid a teacher or learner actually taps through.
Not the Export Dropdown, Not Plugins
Two nearby EduGenius features share vocabulary with this one and are worth ruling out up front:
- Export dropdown (after generation): decides what file your finished content becomes — PDF, Word, PowerPoint, and so on. Covered in full in the export-formats piece.
- Plugins (a separate pipeline entirely): clips web content into a lesson you're editing, with its own honestly-scoped export gap.
The format picker in this piece is neither of those. It's the very first decision in the generation flow: what kind of thing gets made, before any content or file exists to export or clip anything into.
The Real Constraint: A Load Budget, Not a Simple Multi-Select
Every time a format card is tapped, canAddFormatToSelection() runs a full load-budget analysis against the resulting selection — not a static "pick up to N" rule, but a live score built from each format's complexity, credit cost, estimated duration, and requested quantity. This is the picker's most interesting mechanic, and it's genuinely verifiable in lib/generation-complexity-policy.js.
The Hard Ceiling: Three Formats, Always
MAX_GENERATION_SELECTIONS = 3 is a hard constant (lib/generation-complexity-policy.js:8) — no combination of light formats ever unlocks a fourth simultaneous selection. Beyond that ceiling, the load score itself can shrink the allowance further, down to two formats or one, well before a teacher hits three cards.
A cap on simultaneous choices isn't unique to this feature — it echoes a broader finding in choice-architecture research. Sheena Iyengar and Mark Lepper's well-known 2000 study on consumer choice found that a smaller, curated set of options often produces more completed decisions than a large one, because excess choice adds cognitive cost without adding proportional benefit (Iyengar & Lepper, 2000). EduGenius's cap is a technical constraint, not a UX experiment run for that reason — but the same instinct, fewer live options at once, shows up in both places.
The scoring bands themselves are named constants: a "trio-safe" score of 8.5, a "duo-safe" score of 12.5, and a hard limit of 15.5 (GENERATION_LOAD_BANDS). Stay under 8.5 combined and three lighter formats are fine together; cross 12.5 and the picker drops the allowance to two; cross 15.5 and it drops to one, regardless of how many cards are technically tapped.
Two Formats That Always Generate Alone
ALWAYS_SINGLETON_FORMATS hardcodes two format IDs that can never be combined with anything, ever: Long Format Exam and Slide Deck (lib/generation-complexity-policy.js:35). The moment either is selected alongside anything else, the load analysis forces maxSelectableFormats down to 1 and surfaces a blocking reason naming the format.
Here's a nuance worth stating precisely, because it only shows up once you check both files against each other: Slide Deck is hardcoded as an always-alone format, but it's never actually selectable in the picker in the first place. It's tier 3 in config/content-formats.js:429-443, not flagged teacher-only, and not hidden — which means it fails the tier check in both isLearnerGenerateFormat() and isTeacherGenerateFormat() and never appears as a card for either persona. Its singleton rule is real code, just currently unreachable through this screen.
Three Formats With Dynamic Thresholds — And a Catch
Beyond the two hardcoded formats, three more become forced-solo dynamically once a configuration value crosses a threshold, computed inside getFormatSpecificAdjustments():
| Format | Forced solo above | Actually selectable in the picker? |
|---|---|---|
| Presentation Slides | 12+ slides | No — hidden: true excludes it for both personas |
| Image Creations | 3+ images | Yes |
| Data Visualization | 5+ charts | No — tier 3 excludes it for both personas |
Checking config/content-formats.js against this list turns up the same pattern as Slide Deck: two of the three dynamically-gated formats are formats nobody can actually pick from the grid today. Presentation Slides carries hidden: true at line 262; Data Visualization sits at tier 3 with no teacher-only exemption at line 414. Only Image Creations — tier 2, not hidden — is a threshold a teacher or learner can genuinely trigger, by requesting three or more images in one generation.
None of this makes the load-budget engine fake. The scoring logic runs correctly for every format in the registry, including ones the picker doesn't currently expose. It just means the visible, checkable version of "this format forces itself to run alone" is narrower in practice than the code's own constant names suggest: one hardcoded singleton you can actually select (Long Format Exam), one dynamic threshold you can actually trigger (Image Creations at 3+ images), and everything else in that system currently sits behind a filter a teacher never reaches.
What the Picker Actually Shows When You Hit the Wall
None of this is a silent block. When a candidate format would push the selection over budget, its card renders opacity-40, becomes unclickable, and shows the actual reason inline — either the generic ceiling message or the specific singleton warning:
- Ceiling reached: "Maximum formats reached (3). Unselect one to choose another." (
components/OutputFormatSelector.js:327) - Singleton conflict: "[Format name] must be generated alone." (
lib/generation-complexity-policy.js:526, surfaced throughcanAddFormatToSelection())
A teacher who's selected Worksheet and Study Guide, then taps Long Format Exam, sees the exam's card explain itself rather than simply refusing the click — the picker states the reason before a credit is ever spent.
The MCQ Difficulty Gate: A Second, Separate Constraint
Multiple-choice quiz difficulty is capped by requested question count, directly inside the picker's own configuration panel — a second, entirely separate constraint from the load budget above. It has nothing to do with combining formats; it applies even if Multiple-Choice Quiz is the only format selected.
The logic sits inline at components/OutputFormatSelector.js:519-529:
- Under 16 questions: Easy, Medium, and Hard are all available.
- 16 to 25 questions: the Hard option disappears — only Easy and Medium remain.
- 26 or more questions: only Easy remains.
Because the question-count slider steps in increments of 5 (5 to 50, per config/content-formats.js:37), the first value that trips each gate is 20 questions (Easy/Medium only) and 30 questions (Easy only). If a teacher's current difficulty selection falls outside what the new question count allows, the picker auto-corrects it to the first available option rather than leaving an invalid state selected — and shows the reason underneath the dropdown, not just a silently-changed value.
This is a real, checkable UI behavior, not a general "AI gets easier at scale" claim. Selecting exactly 25 questions still allows Hard; selecting 26 does not — the cutoff is a hardcoded number in the component, not a soft heuristic.
Credit Cost Is a Third, Independent Gate
Every format carries its own credit cost, and combining formats sums those costs — but credit affordability and the load-budget cap are two unrelated gates, not one. A teacher can be fully credit-affordable and still selection-blocked by the load budget, or vice versa: able to select three light formats the budget happily allows, while lacking the credits to actually generate them.
Within the 16 formats the picker actually shows, costs run from 2 credits at the cheap end to 24 at the expensive end:
| Cost range | Example formats |
|---|---|
| 2 credits (cheapest) | Multiple-Choice Quiz, Flashcards, Matching Columns, True/False Quiz, Vocabulary List |
| 3–6 credits | Study Guide, Concept Revision Notes, Reading Passage, Rapid Fire Test, Essay Questions, Worksheet, Case Study, Long Format Exam |
| 7–24 credits | Mind Map (7), Image Creations (24 — the priciest format actually selectable in the picker) |
The registry itself lists a 25-credit Slide Deck format (config/content-formats.js:436) — but as established above, that format is never selectable through the picker at all, so 24 credits (Image Creations) is the real ceiling a teacher or learner will ever encounter on this screen.
Combined costs get a flat discount for the Professional plan tier: calculateTotalCost() applies a 10% reduction to the summed subtotal for that plan (CREDIT_RULES.planDiscounts.professional = 0.1 in config/content-formats.js:587), with no discount for other tiers. Selecting two formats simply adds their individual costs together first — there's no separate "combo" pricing beyond that one plan-wide discount.
The Honest Gap: Two of Six Teacher Hero Formats Never Render
The picker's own teacher hero-format list names six formats meant to be featured prominently in the primary grid — but two of them are flagged hidden: true in the format registry, so only four actually appear. This is worth stating as plainly as this site's other feature-mechanics pieces have named their own real gaps: Plugins' clips never reaching the export pipeline, or PDF export's non-English limitation. It's a small, genuine implementation gap, not a broken feature.
TEACHER_HERO_FORMAT_ORDER in components/OutputFormatSelector.js:160-167 lists, in order: Worksheet, Multiple-Choice Quiz, Rubric Grid, Leveled Homework, Presentation Slides, Reading Passage.
Cross-checking each of those six IDs against hidden flags in config/content-formats.js turns up two matches:
- Rubric Grid —
hidden: trueat line 382 - Presentation Slides —
hidden: trueat line 262
The filtering actually happens one step earlier than the hero list itself. allFormats — the full pool both the hero grid and the "+more formats" section draw from — is built from getTeacherGenerateFormatEntries() (components/OutputFormatSelector.js:220-223), and that function excludes anything flagged hidden: true before it even checks teacher-only status.
Rubric Grid and Presentation Slides never enter allFormats in the first place, so heroFormats (order.map((id) => byId.get(id)).filter(Boolean), line 226-230) and tier2 — the "+more formats" toggle — both come up empty for these two IDs. Neither format is reachable anywhere in the picker for a teacher, not just absent from the hero row.
In practice, a teacher opening
/generatesees a primary grid of four hero formats — Worksheet, Multiple-Choice Quiz, Leveled Homework, Reading Passage — where the component's own source names a list of six. Expanding "+more formats" doesn't recover the missing two, either: Rubric Grid and Presentation Slides are excluded from the entire format pool the picker draws from, not filtered out of the hero row alone. There's no error, no placeholder card, and no visible sign either format was ever meant to appear.
Both formats are still real, working format engines elsewhere in the app — the hidden flag only affects this one picker component. But from a teacher sitting at /generate, the gap is total: the six-item hero list names them, and neither the primary grid nor the overflow section ever shows them.
Choosing Formats Well: A Practical Walkthrough
Picture a middle-school teacher building a unit review. They select Worksheet first — the card expands inline to show question-count and question-type checkboxes. Adding Study Guide next is unblocked; the combined load score stays comfortably under the trio-safe band. Adding a third, lighter format like Flashcards still fits.
Now picture the same teacher tapping Long Format Exam instead of Flashcards. The moment that card is selected, every other unselected card dims, and Worksheet's own card (already selected) shows the singleton notice: "Long Format Exam is generated on its own. Unselect it to combine other formats." Nothing crashes or errors — the picker just enforces what the load-budget analysis already decided.
Pro Tips
- Check a format's card for a disabled reason before assuming it's broken. A greyed-out card almost always has a one-line explanation underneath its description — read it before re-clicking.
- Reduce quantity fields to unlock a blocked format, rather than removing a format entirely. Dropping Image Creations from 3 images to 2, for instance, can lift its forced-solo status without losing the format from your selection.
- Treat the "+more formats" toggle as a real catalogue, not overflow. Leveled Homework and other teacher-only artefacts live there permanently, not because they're less finished — the hero grid is a curation choice, not a completeness signal. (Rubric Grid isn't among them — see the honest gap above.)
- Combine cheap, low-complexity formats when you want variety in one pass — Multiple-Choice Quiz, Flashcards, and Matching Columns at 2 credits apiece stay well under every load band together.
- If you rebuild the same format combination weekly, save the whole setup as a Recipe instead of reselecting formats and configuration values from scratch each time.
What to Avoid
- Don't assume three formats is always achievable. The hard ceiling is three, but the live load score can drop that to two or one well before you reach it, especially with a high-risk format like Case Study or Mind Map in the mix.
- Don't confuse the load-budget block with a credit shortfall. A greyed-out card and an insufficient-credit message are two different systems — check which one is actually stopping you before assuming a plan upgrade will fix it.
- Don't expect Presentation Slides or Data Visualization to appear as cards at all. Both exist in the format registry and even have documented singleton logic, but neither is currently reachable through this picker for either persona.
- Don't assume every named "hero" format in the code is visible on screen. The teacher hero list names six; only four render, because two carry a
hiddenflag the ordering array doesn't account for. - Don't build a quiz at exactly 26 questions expecting a Hard option. The difficulty gate is a hard number, not a suggestion — confirm your target question count against the 16/26 thresholds before configuring difficulty.
Key Takeaways
- The format picker is a load-budget system, not a plain multi-select — every selection triggers a live score based on complexity, credit cost, duration, and quantity, computed in
lib/generation-complexity-policy.js. - Three formats is the hard ceiling (
MAX_GENERATION_SELECTIONS), but the live score can lower that to two or one before a teacher ever reaches three cards. - Long Format Exam is the one hardcoded singleton format a teacher can actually select — its sibling, Slide Deck, is also hardcoded as always-alone but isn't selectable in the picker at all, due to its tier.
- Of three dynamically-gated formats (12+ slides, 3+ images, 5+ charts), only Image Creations is genuinely reachable — Presentation Slides is hidden and Data Visualization is excluded by tier, for both personas.
- MCQ difficulty gating is a separate, verifiable rule: 16+ questions drops Hard, 26+ leaves only Easy, hardcoded directly in the picker's configuration panel.
- Credit cost is an independent third gate, ranging 2–24 credits among the picker's 16 visible formats, summed across a combined selection with a flat 10% Professional-plan discount.
- The teacher hero-format list names six formats, but only four render — Rubric Grid and Presentation Slides are filtered out silently because both carry
hidden: truein the format registry.
Frequently Asked Questions
Can a teacher ever select four content formats in one generation?
No. MAX_GENERATION_SELECTIONS is a hard constant set to three in lib/generation-complexity-policy.js, and the live load score frequently caps the real number lower than that, depending on which formats and configuration values are selected.
Why does a format card say it "must be generated alone"?
That message appears when the format is a hardcoded singleton (Long Format Exam) or has crossed a dynamic threshold (Image Creations at 3+ images). The load-budget analysis forces the maximum selectable formats down to one and names the specific format responsible, rather than blocking the click silently.
Does selecting more formats always cost more credits?
Yes, but that's a separate system from the load-budget cap. Credit cost is a straight sum of each selected format's individual cost, with a flat 10% discount for the Professional plan. A selection can be blocked by the load budget while still being affordable, or vice versa.
Why do 26-question MCQ quizzes only offer Easy difficulty?
That's a hardcoded rule inside the picker's configuration panel, unrelated to the load-budget system: 16+ questions removes Hard, and 26+ removes Medium as well, leaving only Easy. It's a fixed threshold in the component, checkable directly in the source.
Are Rubric Grid and Presentation Slides really available to teachers?
Not through this picker. Both carry hidden: true in the format registry, which excludes them from the entire pool of formats the component draws from — not just the hero row. TEACHER_HERO_FORMAT_ORDER still names both, but neither the primary grid nor the "+more formats" toggle ever shows them to a teacher.
Is the 25-credit Slide Deck format ever selectable?
No. Slide Deck is defined in the format registry with a 25-credit cost and is even hardcoded as an always-alone singleton format, but it's tier 3 and not flagged teacher-only, which excludes it from both the learner and teacher catalogues the picker draws from. It exists in code without a path to appear on screen.
Do students see fewer content formats than teachers?
Yes — 15 for learners versus 16 for teachers. The extra format for teachers is Leveled Homework, the only teacher-only artefact in the registry that isn't also flagged hidden.
Related Reading
- How EduGenius's Export Formats Handle Math, Slides, and Language Differently
- How EduGenius's Plugins Feature Clips Content Into a Lesson — But Not Into the Export
- How to Create Coding Quizzes With AI
- How EduGenius's Content History Helps Teachers Find Past Generations Again
- How to Save and Reuse Content Recipes in EduGenius
- How EduGenius's Assign Feature Turns a Generated Quiz Into a Gradable Assignment