Explore the mouth, layer by layer.
A 3D oral anatomy explorer based on a published anatomical atlas. View 265 structures in their anatomical positions or arrange them by category without overlap.
The offline ZIP includes the complete model inside the HTML viewer. Choose Display tools → Download GLB to save the GLB without an internet connection.
D1–D4 learning
Choose D1 · Foundations, D2 · Applied anatomy, D3 · Clinical reasoning or D4 · Comprehensive care in the viewer. English is the default; the language selector also translates the study content into Korean.
- D1: switch between Universal and FDI numbering for the 28 permanent atlas teeth, compare two teeth at the same scale, and read morphology, root-variation and surface-terminology cards. Internal layers opens a separate interactive tooth schematic.
- D2: follow five nerve pathways on either side (the maxillary steps also draw the superior alveolar, palatine and nasopalatine nerves as schematic paths), inspect 35 muscle cards—muscles of mastication, facial expression, the floor of the mouth, the pharynx and the soft palate—whose attachments are painted where the muscle meets bone in this model, and use the Innervation map: choose a nerve or branch (facial nerve and its branches, V3 motor, nerve to mylohyoid, hypoglossal, glossopharyngeal, pharyngeal plexus) to paint the muscles it moves and read what is lost when it is injured. The Anesthesia map covers seven common blocks and five other techniques (local infiltration of one chosen upper tooth, Gow-Gates, Vazirani-Akinosi, AMSA and P-ASA) on either side: it tints the teeth whose pulps a technique numbs and paints the gingiva it reaches on the lip/cheek side or the tongue/palate side. Root & canal cuts any tooth along its long axis, cheek to tongue; for a lower premolar or molar it draws the shortest distance in this model to the inferior alveolar nerve trunk, compares the posterior teeth of that side and gives CBCT values from patients. Explore five step-by-step lessons: local-anesthesia anatomy (including the mental/incisive and infraorbital regions), the five facial nerve branches, how muscle attachments direct the spread of an odontogenic infection, restoration and endodontic anatomy. Jaw motion moves the atlas mandible and lower teeth in 3D with sliders for opening, protrusion and lateral excursion: the first 20 mm open by rotation, wider opening slides the condyles and their discs forward and down, and a card names the muscles at work. The motion is illustrative (typical values; teeth are not collision-tested), and a 2D concept diagram remains below it.
- D3: work through three fictional clinical cases: pulpal/periapical findings, periodontal assessment and a persistent mucosal lesion. Each case reveals information across three stages with choices and feedback. Imaging concepts provides four authored schematics—periapical, bitewing, panoramic and CBCT—with a finding toggle and conceptual slice slider. Periodontal chart lets you edit six sites on Universal #30 / FDI 46: probing depth (PD), signed gingival margin (GM), calculated attachment level (CAL) and bleeding on probing (BOP).
- D4: explore three fictional, three-stage cases on prognosis and phased care, anticoagulant coordination, and replacement options shaped by patient preferences. Care planning has two sequence-ordering exercises. Patient explanations provides three plain-language scripts with teach-back prompts and a text-file export.
- Practice: each year has ten concept questions—40 across D1–D4. Tooth identification draws from 28 atlas teeth; D2 adds Identify muscles, nerves & glands in 3D, which hides the names of four nearby structures of the same kind and asks for one. D2 also generates Neighbours & nerve territories questions from the model’s contact data and the upper-jaw nerve territories, and shows the answer in 3D. Innervation questions, generated from the innervation map, ask what an injury to a branch weakens and which nerve moves a muscle. Which block numbs it, generated from the anesthesia map, paints a site—a tooth’s pulp, or its gingiva on the lip/cheek or tongue/palate side—and asks which block numbs it, or which injections an extraction needs; the answer’s territory is then painted in 3D. Scenario questions ask for the other techniques, such as the closed-mouth block for a patient who cannot open. In D1 and D2, Name the numbered label shows a textbook-style plate (the teeth of one arch, muscles of mastication, floor of the mouth, facial muscles, salivary glands, mandibular and facial nerve branches) with numbered leader lines and asks which structure a number marks; answering adds the names. A session contains up to ten questions, with answers and feedback. Review missed questions in the practice panel. On a phone, questions are answered from a short bar at the bottom of the viewer so the model stays visible; Details and sources opens the full panel.
Practice history stays in this browser’s local storage when available; it is not an account or cross-device record. D3/D4 lab and exercise selections remain in page memory while switching topics or languages; refreshing starts those selections again. The fictional chart is not a saved patient record.
During learning, the study layout controls structure placement and Shift + scroll separation is disabled. Close learning to use normal atlas separation again.
The tooth-layer, jaw-motion, imaging and periodontal illustrations are teaching schematics, not patient scans or calibrated procedure simulations. They are separate from the atlas GLB, are not spatially registered to it, and do not fill its missing anatomy. The static atlas does not acquire the illustrated pathology. Cases and ordering exercises demonstrate reasoning; they are not validated patient simulations or individualized treatment plans.
The notes, diagrams and questions are original study material. Sources & further reading links to supporting ADA, AAE, AAP and academic references; they are not official examination questions or an institution’s endorsed curriculum. Learning scope and source details
Male and female body pages (website)
The website has two more pages, chosen with Region in the header: Male body (2,593 structures of the whole body — the head with the brain, the neck, trunk and all four limbs — from Z-Anatomy, the source of the head and mouth atlas, with the brain from BodyParts3D and the spinal cord from the HRA) and Female body (1,044 structures of the whole body from the Visible Human Female: the HuBMAP Human Reference Atlas 3D reference organs with the mouth, eyes, breasts and intervertebral discs and the Allen brain the HRA places in the same body, the University of Denver musculoskeletal models, the BoneHub CT bone models with the skull, and trunk, neck, jaw and eye muscles, the oesophagus, stomach, thyroid and suprarenal glands, maxillary and frontal sinuses and neck vessels segmented by TotalSegmentator from her CT; the arms are bones only). They use the same viewer: anatomical view, Browse structures with their own categories, presets, layers, sections, measuring, note pins, labels, share links and Download GLB. The D1–D4 learning workspace, schematic paths and university reference models belong to the head and mouth page only. The page files keep their earlier names (pelvis-male.html, pelvis-female.html) so existing links still work.
- Male body (whole body): in the head the skull, ear ossicles and 28 teeth, the jaw joints, the facial, scalp, chewing, tongue and eye muscles, the pituitary; the eyes, lacrimal apparatus, eardrum and auditory tube; the nose, sinuses and nasopharynx; the tongue, gums, soft palate and salivary glands; the cranial nerves, the cerebral and facial arteries, the dural venous sinuses and the facial lymph nodes; in the neck the hyoid, cervical vertebrae, larynx, pharyngeal constrictors, thyroid and submandibular glands, the neck muscles, the carotid, vertebral and jugular vessels, the glossopharyngeal, vagus, accessory and hypoglossal nerves and the cervical nodes; the heart with chambers, valves and papillary muscles, the lungs in lobes, pleura, trachea and bronchi, the thymus; the oesophagus, stomach, duodenum, intestines, liver, gallbladder, pancreas, spleen and omenta; the suprarenal glands and urinary tract; the genital organs; the ribs, sternum, vertebrae, clavicles, scapulae, pelvis and the bones of all four limbs to the fingers and toes with their joints, ligaments, discs, menisci, bursae and tendon sheaths; the diaphragm, chest-wall, abdominal-wall, back, pectoral, shoulder, arm, forearm, hand, pelvic, hip, thigh, leg and foot muscles with their fasciae; the great, coronary, pulmonary, intercostal, abdominal, pelvic and limb vessels to the palmar arches and the foot; the vagus nerves, sympathetic trunks, intercostal nerves, the brachial plexus with the axillary, radial, musculocutaneous, median and ulnar nerves, and the lumbar and sacral plexus; and the lymph nodes. The scalp, the platysma, the fasciae, the chest and abdominal walls, the pleura, the pericardium, the greater omentum and the small intestine start faded so the skull and organs stay visible (Display tools → Layers); the Head & face preset shows the skull, teeth, face and eyes and zooms to the head. The teeth here are the source surfaces; the head and mouth page has higher-resolution teeth and the D1–D4 learning modes. The brain comes from BodyParts3D 4.0 (DBCLS, CC BY 4.0): 74 original surfaces — the gyri and lobes of the cortex, the cerebral white matter, corpus callosum, fornix, hippocampus and amygdala, the caudate nucleus, putamen and globus pallidus, the thalamus, hypothalamus and pineal gland, the ventricles with the choroid plexus, the midbrain and cerebral peduncles, pons, medulla oblongata and cerebellum — in a Brain category and a preset that zooms to it, placed by fitting the BodyParts3D skull onto the page’s skull. Z-Anatomy’s own brain objects and its inner ear are left out because of their licences (no licence stated for the brain models the source adapted; the cochlea and vestibule are non-commercial). Not in the source: wisdom teeth, the intrinsic tongue and palatal muscles, the semicircular canals, the thoracic duct, the anal canal, the perineal muscles and membrane, the scrotum, and the patellar ligament and quadriceps tendon, the cervical plexus and the laryngeal nerves. The pericardium and the phrenic nerves, which neither source has, are drawn as schematic structures fitted to the page's heart, great vessels, anterior scalene and diaphragm; they are marked Schematic in the selection details and hover names, and their shape and course are illustrative, not measured. The spinal cord (30 segments) is the HuBMAP Visible Human Male cord, fitted into this spine level by level and centred in its dural sac. The kidneys are left out because Z-Anatomy adapted them from a non-commercial (CC BY-NC 4.0) model. Two openly licensed kidneys were tested in their place — BodyParts3D 4.0 and the HRA Visible Human Male kidneys — and neither fits this body (15–20 mm from the renal vessels, or running into the psoas, duodenum or suprarenal glands), so none is shown.
- Female body (whole body): in the head the CT skull (cranium with maxillae, and the mandible), the upper and lower teeth, gingiva, palate, tongue with its papillae, cheek lining, floor of the mouth, the three pairs of salivary glands and the palatine tonsils; both eyes in 23 parts each; the brain in 268 regions with the optic chiasm (the Allen atlas brain the HRA places in this body, fitted into the skull — its Brain preset frames it); in the neck the cervical vertebrae, the cervical spinal cord and the cartilages of the larynx; the heart with chambers, septum, papillary muscles and valves, the lungs in their twenty segments with the trachea, bronchi and cartilages, the thymus; the liver with its ligaments, gallbladder, bile and pancreatic ducts, pancreas, duodenum, small intestine and colon, the spleen; the kidneys in layers, ureters and bladder; the uterus, uterine tubes, ovaries and vagina with their ligaments; the ribs and sternum, the thoracic and six lumbar vertebrae with the intervertebral discs, the spinal cord, the hip bones, sacrum and coccyx; both breasts (lobes, ducts, nipple, areola, ligaments and fat); the greater omentum and the renal pelves; the clavicles, scapulae and the bones of both arms and hands; the bones, joint cartilage, menisci and knee ligaments of both lower limbs; the pelvic-wall, hip, thigh and leg muscles; and the aorta, venae cavae and the thoracic, abdominal and pelvic vessels. The lower gingiva and cheek lining, the fat of the breasts, the greater omentum, the small intestine and the ascending and descending colon start faded. Not in the sources: the muscles of the face, jaw, tongue and eye, the nose, ear and cranial nerves (the HRA skin is posed differently from the bones and is left out), the stomach and oesophagus, costal cartilages, diaphragm, pericardium and pleura, the trunk and neck muscles, the hyoid, pharynx and thyroid gland, everything in the arms but the bones (the elbow ends of the arm bones are incomplete in the CT), the external genitalia, urethra, anal canal, pelvic floor, nerves other than the spinal cord, lymph nodes, the iliac and parietal arteries and the vessels of the legs, the fasciae, the foot muscles and ligaments and the leg tendons below the ankle.
- Structure names and descriptions on these pages are in English only for now; the interface follows the language selector. These pages are not in the offline ZIP; after one visit the service worker keeps a page and its model for offline use. The models are large (about 59 MB and 53 MB to download), so they come in parts: the bones appear first and the organs, muscles, vessels and nerves are added as they arrive; on a slow connection the first visit still takes a while before the view can be turned. Details: sources and changes.
Get started
Open oral-atlas.html in a browser. On the website the page appears at once and downloads the model separately (about 16 MB, with progress, kept in the browser cache). After the first visit a service worker keeps the page and the model, so the website reopens at once and also opens without a connection; browsers that support it show Install as app in Display tools. The oral-atlas.html in the offline ZIP contains the model and works offline as a single file. A WebGL-capable browser is required; university reference specimens use the internet. Use the language buttons above, or the viewer’s language selector, to switch between English and Korean.
- Anatomical view shows the assembled mouth. Browse structures arranges it into ten categories. Matching left and right structures appear next to each other at their original scale.
- Choose All categories for an overview, or use the category selector to focus on teeth, muscles, or another group. Within a category, use the previous/next arrow buttons to browse 12 structures per page on desktop or 6 on mobile.
- Find in the header (press / or Ctrl+K) searches structures, display presets and tools from one box: choose a structure to select and frame it, a preset to apply it, or a tool such as section or opacity to open it where it is. On a phone, the search button in the header opens the same box. On a computer, the button at the top left of the 3D view hides the side panel so the model gets the full width, and the Display tools sections fold to their headings; both are remembered. On a phone, drag the handle at the top of the panel to make it taller or shorter (also remembered). Guide and Model details are in the header on a computer and at the foot of Display tools on a phone.
- The toolbar under the 3D view holds the viewer tools in four groups: View (Reset, zoom −/+, Fit), Cut (Section opens the section view), Measure (Distance · angle, and Notes for note pins) and Share (Save image, Copy link, Presentation link). Section and Notes open a small panel over the lower left of the 3D view; Esc or × closes it. The Structure list (phone: Explore) provides Korean and English search; it also accepts the older Sino-Korean names still used in many textbooks, English synonyms and abbreviations (TMJ, IAN) and tooth numbers (#30, FDI 46). Expand Filters for All / Visible and Left / Right filters; collapse them to save space on mobile.
- Select a structure or its name to rotate it in single view. Single view and the 3D views in D1–D4 learning turn freely in any direction, like a trackball, so you can roll a tooth over its crown or root; a view button (Front, Left, Right, Top, 3D) or leaving the view sets it upright again. Return restores the previous browsing position. Expand Histology in the inspector for what the structure’s tissue looks like under the light microscope (in English and Korean, for every structure on all three pages: teeth and mouth, brain, organs and the general tissues), with a labelled schematic drawing for the main tissues (tooth and periodontium, neocortex, cerebellum, alveolus, nephron, liver lobule, intestinal villus, osteon, skeletal muscle, artery and vein, peripheral nerve, retina; Enlarge opens it at full size) and, for 36 notes, a real light micrograph with its author and licence (openly licensed images from Wikimedia Commons; click it to enlarge), and Technical details for triangle counts, dimensions, and source-detail information.
- Open Display tools for presets, surface appearance, labels and the layers: each category can be shown, hidden, faded, or shown Only on its own, and Show all / Hide all switch every category at once (Show all also brings back structures hidden one by one). In the structure list, only hidden structures are marked. Platysma and frontalis are faded by default so the vessels, nerves and ducts beneath them stay visible; clear the option under Layers to show them fully. Labels names up to 14 of the most visible structures (6 on a phone) with leader lines; the names are included in Save image and stay off while learning, except in the numbered label quiz. Use Reset to restore the initial state and Save image to save the current view. Measure has three modes: the straight-line distance between two surface points, the angle at the second of three points, and the shortest distance between two structures you select, with their contact area when they touch; press Esc to finish. Values describe this model, not a patient. Notes on the toolbar attach short notes to points on the model; they appear in Save image and travel with Copy link. Copy link copies a link that reopens the same structure, layout, section and viewpoint—or, while learning, the same topic and lesson step. Presentation link copies the same view for slides or a projector: the link opens with the 3D model filling the window, and Esc or Exit presentation returns.
- Hover over a structure to see its name, including in Anatomical view; on touch screens, tap a schematic path to show its name. In the 3D area, Shift + scroll up increases separation and Shift + scroll down decreases it; the 0–100% slider stays in sync. Up to 60% each structure moves straight outward from the centre of the model and the viewing angle stays, so a little separation shows how neighbouring structures fit together; from 60% the spread structures gather into their categories while the view turns to the front, and 100% shows the category board. Scrolling without Shift still zooms.
- Drag to rotate, scroll to zoom, and right-drag to pan. In the category board, dragging pans the view.
- For keyboard control, focus the 3D viewer: Arrow keys rotate, Shift + Arrow keys pan, + / − zoom, Home fits the view, and Esc closes the open view or panel.
Left and right refer to the person’s body, not the observer. Anatomical view restores the assembled positions. The category board uses a frontal orthographic view; single view allows free rotation. Clipping cuts the structures and draws the inside of each cut structure in a flat colour, so the section reads as filled; it is not a tissue section. Besides the sagittal, coronal and axial planes, Current view · oblique cuts square to the screen: turn the model to look at the cut, then press Align to cut again from the new direction. Copy link keeps the oblique plane.
Edit in Blender
The GLB uses lossless EXT_meshopt_compression. Your viewer or importer must support that extension. In Blender, use a compatible importer; otherwise decode the compression before importing. The bundled browser viewer includes its decoder and works offline.
- Choose File → Import → glTF 2.0 (.glb/.gltf) in Blender.
- Import
oral-anatomy.glb. - Select a structure in the Outliner, then choose View → Frame Selected.
The 265 structures retain their individual names. The GLB includes the procedural tissue textures. Lighting and material appearance may vary between applications. The viewer’s realistic shading for bone, teeth, muscle, soft tissue, glands and joints is generated at display time and is not stored in the GLB. Blender glTF guide
Included structures
| Category | Count | Category | Count |
|---|---|---|---|
| Bones | 11 | Teeth | 28 |
| Soft tissues | 6 | Salivary glands and ducts | 12 |
| Muscles | 72 | Nerves | 46 |
| Arteries | 40 | Veins | 18 |
| Lymphatic tissues | 20 | Joints and ligaments | 12 |
Complete names, descriptions, coordinates, and categories are in structures.json.
Detail improvements and remaining limits
This surface update refines 80 nerve/vessel curves and 20 lymphatic structures, with tissue materials and UV texture coordinates for 104 structures. All 28 teeth keep their previous geometry. Earlier improvements replaced 18 teeth with denser published BodyParts3D surfaces; this update adds both lower first molars, whose outer surface now passes the same registration check (20 teeth, 4.83 times their previous triangle count) and applied the original Blender render-subdivision settings to 17 structures. Subdivision smooths surfaces; display gloss changes how light is shown. These surface refinements and materials do not represent newly measured anatomical detail. This update adds ten structures from the same Z-Anatomy source that passed a registration check: frontalis, the orbital and palpebral parts of orbicularis oculi, platysma and the vagus nerve, on both sides. Levator and tensor veli palatini, on both sides, come from BodyParts3D 4.0 (CC BY 4.0), via the simplified copy published by the human-atlas project. BodyParts3D 4.0 does not share the atlas frame, so one similarity transform was fitted jointly on nine neighbouring structures (sphenoid, maxillae, temporal bones, palatopharyngeus, superior constrictors), which then lie within a median 0.24–0.49 mm of the atlas surfaces (RMS at most 1.23 mm). The complete model contains 265 structures, 1,723,532 triangles, and 922,089 vertices. Read the improvements, measurements, and verification
This is a general adult anatomical model, not an exact copy of an individual’s CT, MRI, or oral scan. It does not fully reproduce every oral tissue.
Wisdom teeth, lip and cheek mucosa and frenula, tongue papillae, taste buds and intrinsic muscles, some oral muscles (styloglossus, palatoglossus and the soft-palate muscles), complete internal tooth layers, the superior alveolar and palatine nerves, the lingual artery, the pterygoid venous plexus, the maxillary sinus, minor salivary glands, and microvasculature are not included. Neither Z-Anatomy nor BodyParts3D 3.0 provides meshes for these. The superior alveolar, palatine and nasopalatine nerves, the lingual artery and the maxillary sinus can be shown as a separate schematic layer (Display tools → Schematic paths) drawn from landmarks in this model; it is not measured anatomy. The tongue retains its original 682-triangle surface; the viewer shows it with curved smoothing through those original vertices (display only, within 0.62 mm of the original faces). The tooth menu separates crown and root surfaces; both lower first molars also retain an original dentine surface.
Surface detail
The viewer opens with All structures selected and the Display tools panel open. The separation slider stays at the lower right of the 3D viewer. Menus, selection details and learning content change inside fixed screen regions. On mobile, Show lesson / Hide lesson lets you read the lesson or inspect the model without leaving your study session.
Realistic tissue is the default display. Bone, enamel and roots, muscle fibers and tendons, tongue papillae, gingival stippling, gland lobules, nerve sheaths, vessel walls and lymph-node capsules each have an illustrative color and texture, and soft contact shading shows where surfaces meet. Choose Atlas colors in Display tools or in the selection panel for distinct category colors. Select a structure, then use Surface close-up to inspect its surface more closely. Nerves remain covered by a sheath rather than exposed braided fibers; no capillary networks, raised lymphoid follicles or papilla geometry are invented.
When a structure is selected, Nearby structures lists the closest structures in this model with their approximate surface distance; select one to move to it. In Inspect individually, those neighbors stay faintly outlined in their anatomical relation. Clear Show faintly in 3D to hide them. With Color contact areas, the surface lying within 0.5 mm of each listed neighbour is painted in the color shown beside that neighbour; point to a neighbour in the list to show only its contact. Distances come from the atlas geometry and are not clinical measurements.
These textures are generated procedurally for visual interpretation, not measured from a new scan. Nerve and vessel routes and branches follow the existing Z-Anatomy atlas. This does not establish patient-specific realism. How surface detail differs from anatomical measurement
University references and sources
University reference models opens Dundee’s CT-and-sculpting-based floor-of-mouth and pharynx model, QMUL’s micro-CT of a real tooth, and UMCG’s educational lymph-node cutaway. These are separate references, not aligned patient data, and have not been merged into the GLB. Their public external viewers require an internet connection.
Base geometry: Z-Anatomy and the GLB distribution. Tooth geometry: the BodyParts3D 3.0_95 STL conversion. Levator and tensor veli palatini: BodyParts3D 4.0, via the simplified copy in human-atlas. The nqwrc application code was not reused.
- BodyParts3D - The Database Center for Life Science - CC-BY-SA 2.1 Japan
- Z-Anatomy - The open source atlas of anatomy - CC-BY-SA 4.0
- Cranial Nerves and Foramina - by University of Dundee, CAHID - CC-BY 4.0
- BodyParts3D, © The Database Center for Life Science licensed under CC Attribution-Share Alike 2.1 Japan
- BodyParts3D, © The Database Center for Life Science licensed under CC Attribution 4.0 International
The adapted base model is provided under CC BY-SA 4.0. Retain attribution, identify changes, and apply the same license to adaptations. External specimens follow their providers’ separate terms. Full attribution and modifications · Original model notice · Original NOTICE · Original full-distribution terms · three.js 0.180.0 · MIT · meshoptimizer 1.3.0 · MIT · fflate 0.8.2 · MIT
Anatomical reference: OpenStax: anatomy and function of the mouth and pharynx
For education and structural exploration. This is not a clinically validated model for diagnosis or treatment.