This page describes how a DomiDo design becomes a step-by-step build guide that a user can follow on a phone, on a tablet, or with a printed reference. DomiDo is built by Avvyland Limited (UK) and sells universal blocks and fasteners only; every construction shown on the platform is a user-generated design, so the assembly experience must adapt to whatever the design happens to be — small or large, simple or complex — without manual authoring. The assembly companion runs on the same 3D viewer described in the 3D and AR page, draws on three different traditions of assembly instruction, and is engineered for the outdoor reality of building from a kit: a phone in one hand, gloves on, sometimes in bright sunlight, sometimes far from a stable network. The companion is deliberately visual and minimal-text so it works across languages and across the range of buyers the platform expects to reach.
Assembly instructions on DomiDo draw on three lineages: the precision and minimal-text style of long-established block toys, the wordless visual style of self-assembly furniture manuals, and the safety and traceability standards of the construction industry. The strongest assembly instructions are visual first, because words add cognitive load and create translation overhead, so the DomiDo style uses a single perspective per step that clearly shows where the new block goes, a callout for each new part with a small parts diagram, a colour highlight on the active step, and minimal text — only when a non-visual cue is essential. Construction-industry conventions apply on larger structural builds, where the instructions follow construction-industry conventions on safety warnings, load-bearing notes, and personal protective equipment (PPE) callouts, aligned with the relevant International Organization for Standardization (ISO) symbols.
Research on assembly-instruction comprehension and self-assembly furniture identifies a small set of recurring error patterns — missed sub-assemblies, incorrect orientation of a part, fastener confusion, and step misordering — and the DomiDo guide is designed to make each of these unlikely. Sub-assemblies are visually grouped and the joining step is marked; orientation cues are baked into the three-dimensional view, and into the universal block geometry itself, where every face is equivalent; fasteners are previewed before they are needed and pre-bagged in the physical kit; and the step order matches the order the platform recommends from the geometry. Academic findings on assembly-instruction effectiveness reinforce the same design choices: interactive three-dimensional and animated guides outperform static images, explode views reduce reorientation time, and progress feedback such as a visible step count, parts remaining, and an estimated time increases completion rates. The assembly companion implements each of these.
The interactive guide is the primary surface and it uses the same 3D viewer described in the 3D and AR page. Each step animates the active block into position with an ease-out curve and a brief emissive pulse, then the camera adjusts to keep the new block centred. Explode view spreads blocks apart from the centre or layer by layer to reveal internal structure, implode view smoothly returns the assembly to its final shape, and combining explode view with playback shows the partial structure expanded so that the next placement is obvious. Future steps appear as faint outlines or ghost blocks so the builder can anticipate what comes next without losing focus on the current step; the opacity is configurable and the default favours minimal distraction. The active block is rendered with a saturated emissive highlight and a thicker edge outline, placed blocks settle to a neutral material, future blocks (when shown) are translucent, and a short caption identifies the block type and its quantity remaining for the next steps.
Rotation, pan, and zoom controls are tuned for one-handed touch use as well as for the desktop. Standard pinch-to-zoom, drag-to-rotate, and two-finger pan are supported, and zoom-to-detail is a separate gesture or button that frames the active block at a useful size for fastener operations. Fastener steps zoom in further than block-placement steps because the components are smaller, and the zoom-to-detail control automatically chooses the right framing for the current step type. A small parts callout panel shows the parts required for the current step — block type, quantity, fastener type, fastener count — and each part is interactive: tapping it brings up an enlarged view. Where the build has natural sub-assemblies, the guide groups consecutive steps into a sub-assembly and shows the user where they are within the larger structure, with a small overview map that highlights the sub-assembly's location in the finished structure.
The platform produces a printable PDF assembly manual alongside the interactive guide, because the manual is useful when the builder wants to leave the phone aside, when they are building outdoors in bright sunlight, or when they prefer paper. The platform renders the same 3D scene into a series of static images for the manual using an orthographic camera for clarity and minimising visual noise, and isometric projection is the canonical choice — it is the most readable angle for orientation, it avoids perspective foreshortening, and it makes parallel lines parallel in the page. Steps are numbered in the same order as in the interactive guide, the numbering aligns with the kit's bill-of-materials sequence so that no extra cross-referencing is needed, and each step page lists the parts required with small inset thumbnails of the block types and fasteners that mirror the interactive parts callout. The first or last page of the manual carries the full bill of materials — every block type and its quantity, every fastener type and its quantity, and the kit identifier that lets the builder verify the package — and safety warnings sit on the inside cover and at any step that requires care (heavy lifts, sharp edges, fastener torque limits, outdoor placement guidance), using standard ISO safety symbols where applicable. The print version optimises for clarity at A4 size and the file is compressed to keep it small enough to share by email or message: text is rendered as text rather than as images, vector graphics are used where possible, and raster images are compressed without losing legibility.
Each step has a clear completion marker: the builder taps to mark the step as done, the guide advances, and the marker remains for a return visit. Step state is stored both on the device and on the server so a builder can switch devices mid-build, and when the builder returns the guide opens at the next incomplete step with the resume entry visible from the home screen, the dashboard, and the listing. Each step carries an estimated time, and the remaining estimate adapts to the actual pace of the build: if the builder is faster than the estimate the remaining time updates, and if slower it adjusts upward. Sections are marked with a difficulty indicator — easy, moderate, or challenging — so that the builder can plan ahead, and a "next three steps" preview helps the builder mentally prepare because many builders prefer to read ahead by a small window. The guide supports undo so a builder can return to a previous step, inspect it, and continue — important for sub-assemblies where a mistake earlier in the build can become visible only after the next step — and a brief completion screen marks the end of the build, offering a clear next action: photograph the build, share it, leave feedback, or return to the gallery.
Building outdoors often means a flaky network, so the assembly companion caches the full assembly data — the 3D model, the texture, the step list, the parts list, the printed manual — on the device the first time it is opened, and a small status indicator shows whether the build is cached. A "download for offline" option lets the builder pre-download everything before leaving the home network, and when the network is unreliable the companion continues using the cached data so that optional features that require the network (such as community comments or share buttons) degrade quietly without blocking the build. When the network returns, the companion syncs progress to the server so that another device can pick up where the build left off.
The error-prevention catalogue covers the most common assembly mistakes — skipped fasteners, wrong orientation of a sub-assembly, wrong block type, and out-of-order steps — and each mistake has a recovery path. Tricky steps carry an explicit visual warning before the user advances, with copy that is short, specific, and dismissable. A small number of steps are irreversible once a fastener is fully torqued, and these require a confirmation tap before the guide marks the step complete. Because DomiDo blocks are designed for reconfiguration, disassembly is a first-class operation: the guide can run in reverse, showing the disassembly sequence and the fastener-removal order. A block-identification guide shows the block types side by side and explains how to tell them apart, which is especially useful before the first build.
Accessibility runs through the entire companion. Colour is supplemented with shape, position, and labels so that colour-blind users do not lose information, and step highlights use both a colour change and an outline change. Step descriptions are accessible to screen readers, and the 3D scene is described textually — which block is being placed, where, how many remain, and what to look out for. Touch targets are sized generously for outdoor use where the user may be wearing gloves and the screen may be smudged or damp, and the companion supports tap and long-press as the primary gestures so that fine-grained gestures such as small drag distances are not required to advance. High-contrast mode is available for sunlit outdoor use, with darker backgrounds, higher-contrast outlines, and brighter active highlights, and voice control allows the builder to advance the step, mark complete, undo, and zoom without putting down the parts in hand.
In a later phase, augmented reality (AR) overlays the next block on the partially built physical structure: the phone shows the next block in its correct position relative to the real blocks, the user picks up the matching physical block and places it. Registration to the physical build-in-progress uses computer vision to recognise the structure and align the overlay, step progression in AR uses the same step sequence as the interactive guide, accuracy and tolerances are tuned to outdoor use, and the companion falls back gracefully when AR tracking is lost. Battery and performance considerations are documented in the 3D and AR page.
The assembly companion can auto-capture a still photograph at each step completion, compile the frames into a time-lapse video at the end of the build, and juxtapose a before-and-after comparison view of the empty space and the finished build. Sharing controls allow the builder to publish the build photo or the time-lapse video, and a build-diary feature retains the time-lapse, the step history, and any notes the builder wrote along the way. The instructions improve over time through several metric channels: assembly-time-estimation algorithms compare actual completion times to the platform's estimates and adjust the estimates accordingly; completion-rate tracking identifies builds that frequently stall and surfaces the troublesome step for review; step-difficulty scoring estimates per-step difficulty from completion data, used to set the difficulty indicators; user feedback per step lets builders flag confusion or errors directly from the guide; and A/B testing of instruction variants tests changes — a different visual style, a different camera angle, different copy — and measures the impact on completion. The assembly-companion technology stack uses the same renderer as the 3D and AR page; the instruction sequence and the time-lapse pipeline run server-side, and the progress state and the user notes are persisted by the backend.
The assembly companion ships in three waves. At launch the companion delivers the interactive 3D guide, the printable PDF, progress tracking, accessibility basics, and the assembly-time estimate. The outdoor-ready wave that follows adds offline caching, sunlit-screen contrast, gloved-hand interaction, and voice control. The intelligence wave further on adds adaptive difficulty, per-step feedback aggregation, the AR assembly overlay, and the time-lapse social features.