Templates & samples

Try the public Signal Conditioning sample, or start from a layout or a lecture: each says the plan it needs, its field, and the level it is pitched at.

Start without an account#

The public filter sample opens one of these lectures without sign-in, on slide 12, “A first-order low-pass”. Increase R while holding C fixed: the cutoff readout falls and the linked resistor value follows. This first task needs no Python. Edits are temporary; reopen the sample to restore its starting values. On a phone the sample opens as a preview with its controls; use a desktop browser for the full editor.

Ten more of the twenty-four lectures double as public demos the same way: six lectures adapted from universities' open courses (each credited on its title and credits slides) and four more of Quodra's own. Open any of them without an account, make the one change each names, and reopen the link to start fresh:

The full gallery and templates below are reached through a Quodra account. They are examples to inspect and adapt; Python output requires enabled execution, trust where requested and a successful run. Only Auto-run cells follow input changes automatically. Canvas and Matter Lab can require separate enablement. To try your own material, start with the one-slide workflow.

Templates#

From a template on the dashboard — or the Template item in the sidebar — starts you from a structural layout built for technical talks rather than a blank slide. A template is a small complete deck, not an empty frame: example prose, cells and a presentation structure you can adapt. Check the content and run the cells you need before presenting. Everything in one is written to be replaced.

Every layout carries two chips on its card. Plan is the lowest plan on which everything in the layout works; all five are Free, because text, code cells, equations, parameters, tables, charts, diagrams and simulations are not gated by plan, and the chip exists so a layout that needs more can say so. Discipline is the field the seeded content is about; Any field marks a layout that suits every subject. Pressing a discipline chip sets the dashboard's tag filter, which narrows the layouts to that discipline plus the ones for any field, and narrows your decks and the gallery to that field at the same time. An AI assistant reads the same two fields from the MCP catalogue (quodra_catalog, kind templates).

Title + Body#

Slide one is a title layout. After it: The claim in a single paragraph, Three points as a bulleted list, a quote slide with one large idea and nothing competing with it, and Takeaways closing on a citation object.

Code + Output#

Built around the idea that the figure is generated, never pasted. Slide two puts a NumPy cell beside its own explanation. Slide three adds two parameter objects, A and tau, so an enabled, trusted Auto-run cell can rerun when they move. Check its status and output before a talk. Slide four is about what to say before showing a figure.

Equation + Note#

Statement, derivation, numerical check, in that order. Gauss's law is typeset as a numbered equation, derived in an aligned block, then evaluated in a Python cell so the closed form is checked against a number — and the deck closes with what the result assumes and what it does not say.

Two-column#

A comparison shape held for the whole deck: the assertion on the left, the thing that makes it true on the right. Slide three seeds a table with a chart bound to it, so editing a cell moves the bars. Slide four pairs a four-node flow diagram with the result beside it.

Problem → Model → Result#

The three-act research talk, worked end to end on a pendulum. The problem slide asks one question a number can answer, beside a three-node flow diagram. The model slide carries the closed-form period, sliders for length and gravity, and a live pendulum simulation. The implementation slide is the cell that reads those same parameters, so its output can follow those parameters after a successful run. The result slide is a measurement table with a scatter chart bound to it, and the conclusion answers the question the deck opened with — nothing else.

Every template arrives as an ordinary deck of yours the moment you pick it, named after the layout. Rename it, delete the slides you do not want, and replace the prose: it was all written to be thrown away.

Samples#

Your library shows a gallery of sample decks under Sample presentations — three complete lectures in each of six fields, every one wearing its field as a tag and a level chip. There are no sections to scroll: click a tag chip, or use the filter bar above the library, to see one field at a time; the level bar beside the gallery's heading, or a level chip on a card, narrows it to one level.

A lecture's level is the level of the course it mirrors; every sample names its source course or textbook in its notes. Three are pitched at graduate level (Maps of the Possible, Cells & Potentials, Ridge & the Fit); the other fifteen are undergraduate. The MCP catalogue (quodra_catalog, kind samples) lists each sample's discipline and level beside its id.

These are not slideware mock-ups. Each is a complete lecture built from real objects — parameters wired into cells, live circuits and simulations, molecules, spectra, beams that actually deflect — so opening one shows what a finished Quodra deck looks like in that field. Every card gives its slide count, and the marker in its corner says whether it was built as a deck or on the canvas.

What is in each one:

Electrical Engineering — Signal Conditioning (schematics that solve on the slide: divider, low-pass, rectifier, a biased common-emitter stage); Poles, Zeros & Stability (a live RC pole you can hold, the s-plane map, Routh's edge found exactly, stability margins); Sampled: Nyquist to Bits (two signals sharing one sample set, a driven anti-alias filter measured by the circuit solver, the price of a bit).

Material Science — Steel, Heat & Structure (the Fe–C diagram with a live lever rule, crystals, simulated diffraction, tensile curves); Maps of the Possible (TTT, Pourbaix, Ellingham, Ashby, an MD vacancy walk, phase-field playback); Diffusion & Kinetics (the interstitial site in the crystal itself, a diffusion movie on a fixed colour scale, carburizing solved exactly, kinetics meeting the TTT map).

Chemistry — Acids & Equilibria (molecules making the inductive argument, an exact bisection titration, an annotated spectrum, the bicarbonate buffer you run on); Rates & Arrhenius (a real 1921 dataset, a fading crystal-violet spectrum family, half-lives from tritium to uranium, catalysis priced in barriers); Cells & Potentials (the potential ladder, the Daniell cell drawn as a cycle, ΔG = −nFE live, Faraday counted in grams).

AI / Machine Learning — Learning by Descent (CS229's own housing data, the update rule converging and diverging, momentum, the bias–variance cliff); Perceptron to Backprop (what one line can settle, the XOR wall, a gradient check with a movable ε, a real 2-2-1 network trained on the slide); Ridge & the Fit (the closed-form estimator, the SVD view, ESL's prostate benchmark, effective degrees of freedom).

Computer Science — The Price of Sorting (one counted sort against theory, recursion trees, the master theorem's three cases, why real sorts are hybrids); Hashes & Collisions (the birthday bound counted across 300 tables, probe costs by load factor, the doubling argument settling at three writes per insert); Shortest Paths (BFS layers measured on a real graph, relaxation as the one move, a counted Dijkstra, why one negative edge breaks greed).

Mechanical Engineering — Bend, Bounce & Break (a cantilever from its governing equation to material selection on a driven Ashby chart, with a live spring-mass sim); Heat into Work (the Carnot ceiling against real reservoir pairs, the Otto loop drawn from its own isentropes, what real engines keep); Flow & Friction (regimes to Colebrook solved live, what a hundred metres of pipe costs, the Moody chart computed).

Opening a sample creates your own copy#

The note beside the gallery heading says it plainly: opening a sample creates your own copy. The samples themselves are built in code and never stored in your library, so nothing you do to one can damage the original, and there is no "reset" to remember.

Opening a sample writes nothing at first. The deck opens with a Sample · read-only chip beside its title; the copy is minted the moment you make your first edit, and from then on it is an ordinary deck of yours — renameable, taggable, filable, deletable. Its field tag comes with it.

Each sample card's kebab menu offers Open a copy and Export JSON, so you can also take one away as a file without adding it to your library.

The public sample behaves differently: edits are temporary and do not become a personal cloud deck. Reopen the original sample link to start fresh. On a phone the sample opens as a preview with its controls; the full editor needs a desktop browser.

Using tags to find them#

The gallery has no categories of its own — only tags, and a level. Every sample carries its field as a tag on the document itself, and the chips on the cards and the filter bar above your library are the whole navigation. All clears the filter; any other chip narrows to that field. The level is not a tag: it lives with the gallery, not in the document, so a copy you make does not carry it, and the level bar beside the gallery's heading is the only place it filters.

Your own decks tag the same way — the card's kebab menu, then Tags…, comma-separated, up to twelve per deck. Because your tags and the sample fields share one filter bar, tagging a deck Chemistry files it alongside the chemistry lectures. Duplicate tags fold case-insensitively, and a tag lives inside the document, so it follows the deck across devices and into every copy.

Blank#

And of course, Blank presentation starts from nothing. See Your first deck for the walk-through, and the Your first technical deck card on Home for the five steps that take a blank deck to a slide with an equation that computes, a chart bound to a table and a circuit that solves.

Next#

• The dashboard — where the gallery, tags, and folders live.
• Your first deck — build from scratch.
• Reactive documents — the idea the samples demonstrate.