Indian Institute of Technology Guwahati
Simhastha 2027
Technology partner
Indian Institute of Technology Guwahati
Simhastha 2027
Technology partner
Presented to the Nashik–Trimbakeshwar Kumbh Mela Authority · Simhastha 2027
Augmented reality on the printed matter of the Simhastha — whatever gets printed, and by whichever office. Augmented reality at the heritage sites of the district. And a network of Experience Centres, each combining virtual reality zones, an immersive theatre, interactive screens and mirror screens. This document explains what each of those is, what it would do for the Simhastha, and at what scale it can be built.
Every route card, darshan calendar, pass, prasad box, hoarding and newspaper page carries a layer that can be opened with a phone. The printed sheet gains a voice, a moving map, a translation, and the ability to correct itself after it has been printed.
What it adds: reach, at national scale, before anyone travels — and it is the only one of the three that occupies no land, needs no staff and forms no queue. Four classes of experience, applied to any printed artefact.
A visitor standing at a temple, a cave or a fort holds up a phone and the place explains itself — what is missing, rebuilt in front of them; what they are looking at; and something they can carry away and repeat.
What it adds: somewhere else worth being. It converts names on a map into destinations, which is the only kind a diverted pilgrim accepts — and it remains valuable long after 2028. 32 sites, around 130 authored points of interest.
Buildings that hold four things at once: virtual reality zones with seated headsets, an immersive theatre needing no device at all, interactive screens to explore at your own pace, and mirror screens that place the visitor inside another century.
What it adds: relief. Sited at the parking nodes, stations, holding areas and medical camps, they give a legitimate reason to be somewhere other than the ghat during the hours the ghat cannot take anyone else. 20 centres, 1,200 seats.
And one thing underneath all three. A single digital twin of the Kumbh geography — the sites, the ghats, the Sadhugram ground — captured once to survey standard. Every AR layer, every film, every screen and every mirror is rendered out of that one record, which is why they agree with each other and why a correction made once propagates everywhere.
These words get used loosely and interchangeably, and they are not the same thing. Each does a different job, suits a different visitor and costs a different amount of space and staff. It is worth being precise about them before any decision is taken, so here is each one described as it would actually be experienced — alongside something comparable that already exists in India.
The visitor holds up a phone. On the screen they see exactly what the camera sees — the temple in front of them, the printed page in their hand — and standing on top of it, in the correct position, something that is not physically there. The tower as it stood before it fell. The name of the carving they are looking at. A voice explaining it in Marathi.
Nothing is built, installed, drilled or attached. The addition exists only on the visitor's screen. Walk away and the site is exactly as it was.
Where this already exists: the Pradhan Mantri Sangrahalaya in Delhi uses augmented reality and smartphone-based exhibits through its galleries. Most people have also met a simple form of it in a phone camera filter.
The visitor sits in a chair and an attendant places a visor over their eyes. The room goes away. They are standing at Kushavarta at five in the morning, before anyone has arrived, and when they turn their head they see what would be beside them. It is not a film being watched. It is a place being stood in.
Eight to twelve minutes, seated throughout, an attendant beside them, and then the visor comes off and they are back in the chair.
Where this already exists: the Anubhuti zone at the Pradhan Mantri Sangrahalaya runs VR experiences for the general public, as do VR darshan counters at Kashi Vishwanath and along the Vaishno Devi yatra route.
A room with a large curved or wrapped screen, rows of seats, and a film of eight to twelve minutes on a continuous cycle. Nothing goes on anyone's face. A family of nine sits together. An eighty-year-old who will not wear a visor watches the same story.
This is where most of the throughput comes from — four shows an hour, every hour the centre is open — and it is the format that excludes nobody.
Where this already exists: the Raigad 4D show and the revolving show at Shivsrushti in Pune; the Shaniwarwada sound-and-light show. The difference here is that ours plays from a computer rather than a fixed reel, so the content can be changed the same afternoon and can run in twelve languages rather than one.
Large touch screens — the size of a door, or laid flat like a table. Nobody narrates. The visitor decides what to look at, in what order, at their own speed, and goes as deep as they want. No queue, no session length, no staff required.
Suited to the things people want to explore rather than be told: all thirty-two heritage sites as a browsable map, the thirteen Akharas and their lineages, the Godavari from Brahmagiri to the sea.
Where this already exists: the Pradhan Mantri Sangrahalaya is built around multi-touch walls and interactive kiosks; most new state museums in India now include them as standard.
A large screen showing the space in front of it, live, the way a mirror does — except that what it shows has been changed. A family standing before it sees themselves on the same ghat, in their own clothes, with the Nashik of eight hundred years ago standing behind them.
The city is what changes, not the person. No face is stored and no individual is altered — a deliberate design decision explained in full later in this document.
Where this already exists: similar installations appear in Indian science centres and at large exhibitions; the form is familiar, the application to a heritage streetscape is not.
Not something a visitor ever sees. It is an exact three-dimensional record of a real place — a temple, a ghat, a cave — made by scanning it, accurate enough to measure from. Think of it as a photograph that can be walked around inside.
It is the master copy. The AR layer, the theatre film, the VR experience and the screens are all produced from it, which is why they never contradict each other. It is also a permanent conservation record of the site as it stood.
Where this already exists: the Archaeological Survey of India and international heritage bodies use scanning of this kind routinely for conservation records of protected monuments.
Each component was chosen because it does something the other two cannot. Taken together they cover a pilgrim from the moment they read about the Simhastha in another state to the moment they show someone at home what they saw.
Component 01 amplifies
The Authority is printing regardless. Augmented reality turns that existing effort into something that speaks, moves, translates and updates. A route card printed in June still gives the correct instruction in August, because the layer behind it can be changed the same morning a ghat closes.
And it is the only component with no footprint at all. It needs no plot, no power, no attendant and no egress plan. It happens in a queue, on a train, in a tent, and in a household in Nanded a month before anyone travels.
Component 02 amplifies
Part of the pressure on the main ghats is simply that people have nowhere else to be. A site with nothing at it is a name on a map; a site that explains itself is a destination. Thirty-two interpreted sites give the Authority somewhere real to send people.
And unlike everything else in this proposal, this component does not end. The layer stays live at every site after July 2028, which converts a twenty-one month event into permanent tourism infrastructure for the district.
Component 03 amplifies
Crowd managers do not need people to move faster. They need people to accept a hold. The traffic plan already stops private vehicles several kilometres out and moves pilgrims onto shuttles — which means captive waiting is a designed feature of the plan, not an accident.
An Experience Centre at that point turns a wait into something received. And four formats under one roof means nobody is excluded — the visitor who will not wear a visor still gets the theatre, the screens and the mirror.
And they amplify each other. The print seeds the heritage visits — a family decides to see Gondeshwar while reading a supplement in Pune. The Experience Centres carry the famous sites that the heritage layer deliberately leaves out, so Ramkund gets more interpretation, not less. And all three are drawn from the same digital twin, so a pilgrim who meets the Simhastha on a newspaper page, then at a cave, then in a recliner, is being told one consistent story rather than three unrelated ones.
Every figure below is a delivery quantity, arrived at by counting real sites, real artefacts and real seats. Each is broken down in full in the sections that follow, and each can be scaled by the Authority in defined tiers.
Augmented reality can be delivered two ways. Through a web link, which needs no installation but needs signal at the moment of use. Or inside an application, which downloads its content once and then works with no connection at all. At a Kumbh, only the second is honest.
/01
Site content is fetched over Wi-Fi — at the station, at an Experience Centre, at the tent city, or at home before travel — and cached on the handset. From that point the layer runs in aeroplane mode. A pilgrim in Panchavati with no signal gets the full experience.
/02
A pilgrim will install one government application for this Mela, not two. Everything here rides inside the Authority's app as a module its own development team integrates — not as a competing download the public has to be persuaded to find.
/03
Every AR experience and printed artefact resolves back into KumbhDoot for anything it cannot answer — the same assistant, helpline and lost-and-found, across WhatsApp, IVR, kiosk and app. The pilgrim never crosses a brand boundary.
What this means practically. We deliver a documented software module and content pipeline to whoever is building the Authority's application. Integration is theirs, content and experience are ours, and the Authority owns both. A lightweight web route is retained as a fallback for a visitor who will not install anything, but it is the exception rather than the design. We would ask to be brought into the application's architecture conversation early, because offline caching at this volume is a decision that must be designed in rather than added afterwards.
The Authority and its departments will produce printed matter across many offices over twenty-one months, and no one can list all of it today. That does not matter, because the layer is not tied to a particular artefact. Anything printed can carry it — a route card, a calendar, a prasad box, a hoarding, a newspaper page — and what the layer then delivers falls into four classes. The component includes the print itself, because a marker added to somebody else's finished artwork performs a fraction as well as one designed into the composition from the start.
| Class of experience | Printed matter it applies to | What the layer does |
|---|---|---|
| Class 01Wayfinding & live maps | Official Kumbh map · route and marg card · parking and shuttle pass · ghat allocation slip · Sadhugram and Akhara directory map | The printed map becomes a live one: current ghat status, diverted margs, shuttle positions and the walk from where the visitor is actually standing. A card printed in June still gives the correct instruction in August. |
| Class 02Instruction & advisory | Darshan calendar for the Nashik and Trimbakeshwar muhurtas · health, heat and water advisory · do's and don'ts card · volunteer and marshal handbook · tent city and dharamshala welcome kit | Spoken instruction in the pilgrim's own language, for a reader who may not read at all. The marshal's handbook shows the plan as it stands this morning rather than as it stood at print. |
| Class 03Devotional & ritual | Aarti and stotra booklet · panchang · prasad packaging and sealed boxes · snan commemorative folio | Sung aarti with the words following, the meaning of the muhurta being observed, and the story of the tirtha the pilgrim is standing in. The devotional object becomes the doorway. |
| Class 04Storytelling & invitation | Newspaper supplements in Marathi, Hindi and English · hoardings on the Mumbai–Agra corridor, ring road, station and airport approaches · rail and MSRTC ticket jackets · MTDC and Maharashtra Tourism collateral | A preview of the thirty-two heritage sites, delivered before travel and outside the district. This is where dispersal is seeded — a family decides to visit Gondeshwar while reading a supplement at home. |
And print becomes measurable for the first time. Opens logged by artefact, placement, page, language, hour and district of origin. After the first Amrit Snan the Authority knows which pieces are being used and which are being discarded, and can print differently for the second. No administration has previously had attribution on Kumbh print.
Ramkund, Kushavarta, the Kalaram approach, Sita Gufa, the Trimbakeshwar sanctum and the ghat faces are not AR sites. They are where density is the operational problem, where a stationary visitor holding a phone is an obstruction, and where the crowd plan is actively trying to keep people moving. Those places are served instead inside the Experience Centres — explained at length, from a seat, several kilometres away. They receive more interpretation this way, not less. What remains is a different and better inventory: thirty-two sites of genuine worth where a visitor can actually stop.
Within the city, reachable on foot or by a short auto ride, away from the bathing corridor.
The second Kumbh centre has a hinterland almost nobody interprets.
The tier that gives a diverted pilgrim somewhere worth going. Several sit on the Nashik–Shirdi corridor.
Group C is confirmed by joint survey with the Tourism Department and MTDC; substitutions within the tier are expected and welcome.
Most heritage technology does only the middle one — it puts text on a screen and calls that interpretation. Every point of interest in this programme is authored across all three layers, and a point is not signed off until all three exist.
The reveal. What is missing, rebuilt and standing in place. The tower restored to its original height above the surviving base. The light of a different hour. A carved narrative band moving through its sequence. Scale shown against a human figure. Nothing on this layer is text. It exists to buy the eight seconds in which the other two become possible.
What it is, who made it, when, under whose patronage, in what material and why that material dictated the form. Delivered as audio first with visual support, at the pace of someone standing in front of the thing rather than reading about it later. Attached to the specific feature, not to the site as a whole.
A comparison that makes the knowledge portable: this against that, this period against the next. A "read the stone" mode that teaches a visitor to identify a motif and then find three more unaided. A trail connecting this site to the next. And a short takeaway into their app, in their language, to show someone who was not there.
Why the third layer matters most. Engagement produces a photograph. Information produces a nod. Learning produces the sentence a pilgrim says at home three weeks later — which is the only one of the three that brings anybody back to Nashik outside a Kumbh year. It is also the only layer that can be tested with an instrument rather than a headcount, and it is what the evaluation programme is built to measure.
An eleventh- to twelfth-century Shiva temple of the Yadava period in black basalt, on a platform of roughly 125 by 95 feet. A complete panchayatana: a central shrine with four subsidiary shrines at the corners for Vishnu, Surya, Parvati and Ganesha. Bhumija shikhara. Widely held to be the largest, most complete and best preserved medieval temple in the Deccan — and on most days almost empty. Twenty-six kilometres from Nashik on the Shirdi road, which is to say directly on a corridor a diverted pilgrim is already travelling.
The same three-layer treatment is authored for each of the thirty-two sites, at an average of four points of interest per site — roughly one hundred and thirty authored points in total. The exact count is fixed by survey and reported to the Authority before production begins.
A headset suits some visitors and not others. An eighty-year-old will decline one; a family of nine wants to sit together; a child wants to touch something. So each centre carries the four formats under one roof — VR zone, immersive theatre, interactive screens and, at the larger sites, a mirror screen — and the visitor chooses their own way in. Every location sits where the crowd plan already wants people: at the parking nodes where private vehicles are stopped several kilometres out, at the shuttle boarding points, the stations, the tent city and the medical camps.
Where
Tapovan / Sadhugram precinct, Nashikand Ahilya Kund / permanent Sadhugram, Trimbakeshwar
VR zone
60 reclinersHosted, seated, 8–12 minutes, reset between guests
Immersive theatre
120 seatsWrapped screen, shared viewing, no device required
Interactive screens
8 stationsTouch walls and low tables for standing and seated use
Mirror screen
1 room-scale installationThe Mirror of the Marg — described below
Alongside
4 interactive AI guide kiosks · live darshan and ghat status board · 2 keepsake stations · quiet and accessible bay
Where — Nashik
Nashik Road railway station · Nashik CBS · Mahamarg bus stand · Tapovan tent city · Kala Gram / Cultural Haat · two satellite parking and shuttle-boarding nodes · Nashik Civil Hospital and main medical camp
Where — Trimbakeshwar & region
Trimbak road holding area on the Nashik–Trimbak corridor · Trimbakeshwar bus stand and parking · Ozar airport, following the expansion · Shirdi, which will absorb overflow whether or not it is planned for
VR zone
24 recliners each288 across the twelve
Compact theatre
40 seats each480 across the twelve
Interactive screens
4 stations each48 across the twelve
Alongside
1 AI guide kiosk · live darshan and ghat status board · accessible bayThe status board alone makes the centre worth entering: it tells a pilgrim when it is sensible to move
Purpose
The Authority's flexible reserveRelocatable within a single shift, on the crowd team's instruction, to wherever holds are actually forming
VR zone
12 recliners each72 across the six
Indicative deployment
Brahmagiri route · Anjaneri · Igatpuri · Sinnar · Ghoti · Vani · and the shuttle-boarding areas on Parva days
Power and link
Self-powered, satellite or bonded-cellular backhaulDesigned to operate where the fibre network does not reach
Network capacity, and the limit we will state before anyone asks
And the honest limit: measured against the Authority's own peak-day projection, this network still serves a fraction of one per cent on an Amrit Snan morning. It is not a headcount instrument on those three days and anyone presenting it as one is selling. Its value lies across the other six hundred and thirty days of a twenty-one month Mela, at the nodes where holds form, and in the specific cases — the elderly, the ill, the pilgrim who travelled a thousand kilometres and cannot make the final walk — where it is the difference between a completed pilgrimage and an abandoned one. On peak mornings we plan for closure unless the crowd team wants a node open as holding relief.
Format 01
Recliners in rows, an attendant to every four seats, headsets sanitised and reset between guests. The deepest of the four — and the slowest, and the one some visitors will refuse. Eight to twelve minutes, tap to start, one physical button for reset. Carries the places nobody can otherwise reach: Kushavarta before dawn, the inside of Sita Gufa, the sanctum approach.
Format 02
Where most of the people actually go. Nothing on anyone's face, a family sits together, and four shows run every hour. This carries the story pieces — the Simhastha itself, the Godavari, the Peshwai — and it is the reason an elderly pilgrim is not excluded from the centre their grandchildren dragged them to.
Format 03
For the things people want to poke at rather than be told. All thirty-two heritage sites as a browsable map with routes and distances. The thirteen Akharas and their lineages. The Godavari from Brahmagiri to the sea. The temple-dating comparison from the learning layer. Self-paced, no queue, no staff, no session length.
Format 04
The Mirror of the Marg, at four locations — both anchors, Kala Gram and the Nashik Road concourse. The space in front of it rendered into the Nashik of another century: the same street, the same ghat, the same procession route, eight hundred years apart. Visitors see themselves in the frame as a crowd, in their own clothes.
Alongside
KumbhDoot given a screen and a voice, for the pilgrim with no smartphone and nobody to ask. Twenty kiosks across the network, at standing and seated height, voice-operable end to end, in the same language set the Authority has already standardised on.
Alongside
A live darshan and ghat status board at every centre — the reason a pilgrim walks in even if they want nothing else, because it tells them when it is sensible to move. And a keepsake station that prints a commemorative card carrying its own layer, so the visit continues after the visit.
On the mirror, a design decision worth stating openly. The obvious version of this installation dresses the individual visitor in period or religious attire. We have not proposed that version. At a Kumbh, an installation that alters a devotee's likeness into borrowed religious dress carries reputational exposure for the Authority and no corresponding benefit, and it requires face capture on pilgrims that this programme rules out elsewhere. The version proposed alters the city, not the person. No face is stored and no individual is modified.
A live, very high resolution, 360-degree relay from fixed positions at Ramkund and Kushavarta, carried to every centre in the network and played in both the VR zone and the theatre. The pilgrim is placed at a vantage point the physical queue cannot offer anybody: still, unobstructed, at the front, for as long as they wish to stand there. Not a broadcast on a screen — a position.
Capture
Fixed 8K 360° camera rigs, two to three per siteProfessional stitching servers, sub-second latency. Positions agreed with police and fixed in advance; no operator moves inside the cordon.
Why 8K and not 4K
A headset shows only 90–120° of the sphere at a timeAt 4K the visible portion is soft. 8K is the practical floor for this to read as presence rather than as video.
Bandwidth
Order of 80–150 Mbps per feedNot deliverable over public mobile networks on a Parva morning. Entirely deliverable over the Authority's own 385 km fibre — which is precisely what makes this possible in Nashik and impossible at most events.
Delivery
Viewport-adaptive streaming to the centresFull resolution sent only to where each viewer is looking, which keeps the aggregate load across twenty sites manageable.
Naming
To be ratified, not assumedWhether this is presented as darshan or as a relay is a question for the Akhara Parishad and the Trusts, not for us. The build is identical either way; the words on the screen are not.
Treatment in scope
A separate, switchable componentBuilt so it can be withdrawn in June 2027 without disturbing any other part of the programme.
Three approvals gate this, and we will not pretend otherwise. Camera positions inside a police cordon on the highest-pressure morning of the year. Guaranteed backhaul on the Authority's fibre, contracted rather than assumed. And a ruling from the Akhara Parishad on whether a live relay of the snan may be presented at all — because if the answer is that it is a broadcast and not a darshan, it must be named a broadcast. If any of the three does not come, the centres carry the recorded library instead and nothing else in the programme changes.
Eight to twelve minutes each, by design: long enough to build the memory of having been somewhere, short enough that nobody starts working out how to leave politely. Every script is developed with the Akhara Parishad, the relevant Devasthan Trust and the state's own historians, and approved before a line is recorded.
| # | Experience | Format | What it carries |
|---|---|---|---|
| 01 | Simhastha — why here, why now, why every twelve years | Theatre · 12 min | Samudra Manthan, the falling drops, Jupiter's entry into Simha, the Godavari's descent from Brahmagiri. Most pilgrims can name the ritual and not the cause. |
| 02 | Godavari — from Brahmagiri to the sea | Theatre · 9 min | The river as the reason for everything else. Source, course, and why the two towns sit where they sit. |
| 03 | Kushavarta at dawn | VR · 8 min | The kund empty and still, at an hour no pilgrim will see during the Mela. |
| 04 | Trimbakeshwar — the approach | VR · 8 min | The Jyotirlinga approach and carving programme at a proximity the physical queue cannot allow. Subject to Devasthan Trust approval on what may be shown. |
| 05 | Ramkund and the ghats through the centuries | VR · 8 min | 1696, Gopikabai's restoration, and the ghat sequence as it changed. The site the pilgrim is queueing for, explained while they queue. |
| 06 | Panchavati — the Ramayana in place | VR · 8 min | The five banyans, Sita Gufa, Tapovan and the episode that named the city — including the interior of Sita Gufa, which most visitors physically cannot enter. |
| 07 | Pandavleni — two thousand years of carving | VR · 8 min | The Buddhist excavations on Trirasmi hill and their inscriptions. A dispersal instrument as much as an experience: it makes people want to go. |
| 08 | The Peshwai | VR · 8 min | The Akhara procession, from archival material and new capture, with the Akharas' participation and consent. For everyone who will never get near the route. |
| 09 | Nashik and the first Indian film | Theatre · 7 min | Phalke, Raja Harishchandra, 1913 — and the hillside where the same stories were carved two millennia earlier. |
| 10 | Gondeshwar and the Yadava century | Theatre · 9 min | Growth slate. The temple restored, and the district trail that follows it. Seeds the Group C dispersal programme from inside the centre. |
| 11 | Anjaneri | VR · 8 min | Growth slate. A destination most pilgrims never reach, and a hill worth diverting to. |
| 12 | Saptashrungi Gad | VR · 8 min | Growth slate. A great Shakti site of Maharashtra, sixty kilometres out, and a genuine pressure-release destination. |
| 13 | Nivruttinath and the Nath sampradaya | VR · 8 min | Growth slate. Central to the Trimbakeshwar pilgrimage and almost entirely uninterpreted for visitors. |
| 14 | The live relay | Live · VR + theatre | Approval-gated and wholly separable. Plays to both the VR zone and the theatre, so a family who will not wear a headset still receives it. |
Languages
Marathi, Hindi and English from Dhwajarohan; twelve by the Amrit Snans; extending toward the full BHASHINI set already standardised for KumbhDoot. Fourteen experiences across twelve languages is 168 discrete audio deliverables — the largest single production line in the programme.
Production
A significant share of pilgrims will not read a screen. Every experience is authored so the audio alone carries the full content, with the visual as support. This also makes the audio-described version a by-product of the main build rather than a bolt-on.
Refresh
A finished film is finished. Across twenty-one months the Authority will want to change a name, a date, a route, a piece of guidance. In a real-time build most of that is an afternoon's work by the Authority's own team — which is what stops the network going stale by the third quarter.
The thirty-two heritage sites, the ghat sequence, the Sadhugram ground and the centre locations are recorded once, to survey standard, and assembled into a single georeferenced twin. The AR layers, the theatre films, the VR experiences, the mirror screens and the interactive screens are all produced from that one record — which is why they agree with each other and why a correction made once propagates everywhere.
It also has a life independent of this programme: a conservation baseline, a crowd-simulation asset for the command centre, and the planning reference for 2039.
Method
Splat capture holds carved basalt, wet ghat steps, brass, water and low interior light in ways older photographic methods do not, and renders on a handset rather than a workstation. Where dimensional accuracy is required — conservation record, crowd modelling, structural reference — it is registered against a laser scan. Non-contact throughout; nothing is touched and nothing is attached.
Window
From early 2027 the sites are scaffolded, hoarded or occupied, and the Simhastha then runs to July 2028. Sixteen weeks, two teams in parallel — one per town — averaging three sites per team-week including permissions, night access and reshoots.
Precedent held by the technology partner
Every wall, stair and courtyard of a fortress complex, captured end to end. The closest existing analogue to what this programme requires, and demonstrable on a screen rather than described.
Dhundiraj Govind Phalke made Raja Harishchandra and invented a form. His memorial stands at the foot of Pandavleni, below caves where people were carving the same stories into rock two thousand years earlier. Nashik is not adopting immersive storytelling in 2027. Nashik has been the place where sacred narrative meets the newest available machine for a very long time, and the machine has changed perhaps four times.
Why it is not merely a good line
Somebody will say that headsets and screens cheapen a religious occasion. It is a fair objection and deserves a real answer rather than a defensive one. The answer is on that hillside: rock, then paint, then Phalke's camera, then this. The form has always changed. The obligation — that it be done reverently, accurately and well — has not.
What it gives the Authority
Prayagraj has the Sangam. Haridwar has the Ganga. Only Nashik can claim to be where India first told these stories with a new machine — and the site that proves it is a Group A heritage site in this programme, a hundred metres from the memorial.
Nothing here requires a new exemption. The purpose of this section is to show that the statutory route for each component has been identified in advance, so approvals are a schedule rather than a discovery.
These are drafted as contractual restrictions rather than as a statement of values. If the Authority wishes any relaxed, that is the Authority's decision to record in writing. The default is no.
The Authority has set the standard with KumbhDoot: a pilgrim on a basic feature phone must receive the same service as one on a smartphone. The in-app layer cannot meet that standard alone, and we will not claim otherwise. The three components are designed to cover for each other.
Prayagraj deployed at scale and published visitor counts. That is not the same as evidence. This programme is led by an institute rather than a supplier precisely so the measurement design is independent of the people building the installations, and so the findings are published whichever way they fall.
Layer 01
In-app opens and sessions by site, artefact, placement, language, hour and district of origin. Occupancy, queue length, refusals and drop-outs at all twenty centres, by format. Visitor counts at all thirty-two heritage sites, week over week, against a baseline set before the Mela opens.
Layer 02
The question that matters. Did Group C sites gain visitors, and did the ghats lose pressure in the same window. Dwell at nodes with a centre against matched nodes without one. Willingness-to-hold, recorded by the marshals working those points. Sourced from the Authority's own command-centre data, not ours.
Layer 03
Structured intercept surveys in Marathi and Hindi at centre exits and heritage sites, against a matched sample of non-users. Did they retain what the learning layer taught. Would they return outside a Kumbh year. Did it feel respectful — asked directly, and reported even when the answer is unwelcome.
What the institute contributes that a supplier cannot. Baselines set before deployment rather than after. Instruments designed by people with no commercial interest in the result. Peer-reviewed publication, which makes Nashik the reference edition that Ujjain, Haridwar and Prayagraj cite. Capacity building, so Authority staff and Maharashtra institutions can operate and extend the system beyond 2028. And an assurance layer over delivery the Authority does not have to construct itself.
Sequenced so the irreversible work — capture — happens first, and so the Authority holds a genuine scaling decision at the end of each phase rather than a fait accompli.
Joint survey with the Tourism Department and MTDC fixing the final thirty-two sites and the exact point-of-interest count. Permissions convened under Section 8: Devasthan Trusts, ASI, police siting protocol. Data Protection Impact Assessment delivered. Baselines set with the Authority's data team. And the integration conversation with the Authority's application team, because offline caching at this volume must be designed in rather than added later.
Sixteen weeks, two parallel teams: all thirty-two sites plus the ghat sequence and Sadhugram ground captured and assembled into the twin. In parallel, all twenty-four print families designed, produced and live, with the administration panel handed to the Authority's communications cell. All fourteen scripts written and approved.
Two anchor and twelve node centres fabricated, sited and commissioned; six mobile units delivered to the crowd team. All heritage layers live in the app. Nine experiences finished in twelve languages. Interactive walls, mirror screens, kiosks and status boards installed and connected to the command centre. Attendant training with Authority staff. Two full-load dry runs with the crowd team empowered to fail the programme. Relay go/no-go in June.
Full operation, with the delivery team on site in support rather than in command. Daily readouts to the command centre. Centres closed on peak mornings unless the crowd team wants a node open as holding relief. The first independent data on immersive provision at a Kumbh is generated here.
Fourteen months of Simhastha remain after the Amrit Snans, at manageable density — when the centres perform best and when the growth slate comes online. Language coverage extends toward the full BHASHINI set. Dispersal is measured properly across a full season.
The digital twin and the thirty-two-site library, owned outright. The heritage layer live in perpetuity. Anchor centres relocating to the permanent Sadhugram and the district's museums. The evaluation dataset published. And the conservation baseline and planning reference for the 2039 cycle, made at the only moment it could have been made.
Institutional lead
Engaged under a Memorandum of Understanding as institutional and technical lead: programme governance, independent evaluation design, technical assurance over delivery, capacity building for Authority and state staff, and publication of findings. Principal Investigator: Prof. (Dr.) Debayan Dhar, Department of Design, IIT Guwahati.
Technology partner
Contracted beneath the institutional lead for design and delivery: capture and the digital twin, the in-app AR module, interactive print design and production, real-time experience builds, Experience Centre design and commissioning, and operator training. Relevant record includes a complete site twin of the Golconda Fort complex delivered in around two weeks, and immersive training delivery for the Indian Army completed ahead of schedule.
Named partners
Structures, optics, compute, screens and power through named partners, itemised separately and disclosed before commitment. Attendant staffing recommended through Authority deployment or a local self-help group rather than an agency — it costs less and places Mela income in Nashik households.
Two limitations recorded up front. Volume 3D and asset production runs through a vetted specialist network that the technology partner art-directs and whose files it owns — accountability sits in one place, but this is not a large in-house production department. And no organisation in this partnership has delivered at a Kumbh before; very few anywhere have. The phasing is constructed so that this is tested against real conditions in Phase 0 rather than discovered in August 2027.
Augmented reality adds something to the real place you are standing in, seen through your own phone — you are still at the temple. Virtual reality replaces the real place entirely with another one, seen through a visor — you are no longer in the room. AR is for people who are already at a site. VR is for people who cannot get to one.
Because augmented reality requires a person to stand still and hold up a phone, and Ramkund is the one place in Nashik where the Authority is actively trying to prevent exactly that. A stationary visitor on a ghat approach is an obstruction, and on a Parva morning it is a hazard. Ramkund is instead one of the richest pieces in the Experience Centre library — explained at length, from a seat, several kilometres away, while the pilgrim waits for a window to open. The site gets more interpretation this way, not less.
Because a web link needs signal at the moment of use, and on the days this matters the network in Panchavati and Trimbakeshwar will not carry it. An in-app module caches its content once over Wi-Fi and then runs with no connection at all. It also keeps the pilgrim inside the Authority's own environment, with its identity, language set and assistant, rather than sending them to a web page. A browser route is retained as a fallback, but it is the exception rather than the design.
The correct first question, and the programme is built around it. Nothing is designed until police and the crowd-modelling team have approved a location. Every centre carries a stated maximum standing density, a documented queue geometry and egress width, and a named officer empowered to close it. The default on Amrit Snan mornings is closure unless the crowd team wants a specific node open. The centres are sited at the parking and shuttle nodes where vehicles are already being stopped — where captive waiting is a designed feature of the traffic plan, not an accident.
Done badly, yes. Three things make it so and all three are ruled out contractually: presenting a screen as a substitute for the snan, generating rather than recording sacred imagery, and monetising darshan. Every script is approved by the Akhara Parishad and the relevant Devasthan Trust before recording, and where a Trust declines, the site leaves the programme. Even the naming of the live relay is put to the Parishad rather than decided by us. The affirmative argument is the Pandavleni hillside: rock, then paint, then Phalke's camera, then this.
Because "almost nobody visits" is the point. The dispersal value of the heritage component comes entirely from the district ring, and it only becomes a real alternative destination once there is something at each site to encounter. Interpreting only the famous sites concentrates visitors on the two places already under most pressure. Gondeshwar is also, on the architectural merits, among the finest medieval temples in the Deccan, sitting free and empty twenty-six kilometres from Nashik on the Shirdi road. It costs the Authority nothing in crowd management and returns a genuine destination.
A different layer entirely. What has been convened is a sensing, prediction and assistance stack, and none of it produces anything for a pilgrim to look at. Somebody still has to capture Kushavarta at first light, get a Marathi script past the Akhara Parishad, fabricate a structure that survives a Tapovan monsoon, and train an attendant who has already run nine sessions that morning. This programme sits underneath that stack and consumes its outputs. Access to the Kumbh AI Sandbox for stress-testing would be valuable and is requested.
Each component is separable with a defined fallback. If a Devasthan Trust declines, that site drops and the count reduces. If the Akhara Parishad rules against the relay, the centres run the recorded library. If police decline a node, that node is not built. If the application team cannot integrate in time, the browser fallback carries the heritage layer at reduced quality. Nothing is architected so a single refusal takes down the rest — a deliberate choice, and the reason the approvals path is scheduled in Phase 0 rather than assumed.
It belongs to the Authority. The digital twin, the capture library, the experiences, the AR module and the evaluation dataset are owned outright; interactive builds and the master source are licensed to the programme with ownership, hosting and hand-off recorded in writing before any payment. In practice: the heritage layer stays live at all thirty-two sites permanently, the anchor centres relocate to the permanent Sadhugram and the district museums, and the twin becomes the conservation baseline and the reference for 2039.
Augmented reality on the printed material, which reaches a household in another state before anyone boards a train and costs the Authority no land at all. Augmented reality at thirty-two heritage sites, which gives a diverted pilgrim somewhere worth going and keeps working long after the Mela closes. And twenty Experience Centres carrying twelve hundred seats, placed exactly where the traffic plan is already holding people, with four different ways in so that nobody is excluded.
All three draw on one digital twin of the Kumbh geography, and that twin can only be made in the window between the works finishing and the crowds arriving. Whatever is concluded about the three components, the record of Nashik and Trimbakeshwar as they stand in the winter of 2026–27 is worth making for its own sake — and the next opportunity is 2039.
Prepared for the Nashik–Trimbakeshwar Kumbh Mela Authority, September 2026. Site inventories draw on the Nashik District Administration's published register of monuments, Maharashtra Tourism and Incredible India records, and Archaeological Survey of India listings; Group C is subject to confirmation by joint survey. Comparable installations referenced for illustration — the Pradhan Mantri Sangrahalaya, Shivsrushti at Pune, the Shaniwarwada sound-and-light show, and VR darshan counters at Kashi Vishwanath and Vaishno Devi — are the work of other organisations and are cited only to describe the formats. Statutory references are to the Nashik–Trimbakeshwar Kumbh Mela Authority Act, 2025 (Act 33 of 2025); the Ancient Monuments and Archaeological Sites and Remains Act, 1958; the Rights of Persons with Disabilities Act, 2016 and the Harmonised Guidelines and Space Standards for Universal Accessibility in India, 2021; IS 17802 as applied through the RPwD (Amendment) Rules, 2023; and the Digital Personal Data Protection Act, 2023. The digital backbone and assistant described are from NTKMA public statements and briefings, 2025–26. Point-of-interest counts, session throughput and derating factors are planning assumptions derived from stated capacity and are labelled as such wherever used; they are not forecasts, and delivery quantities describe the proposal's scope rather than an estimate of demand.