Best Augmented Reality Companies in San Francisco

Quick answer: San Francisco is the rare market with a genuine local augmented reality industry, but it serves consumer brands, retail and enterprise rather than property development. Hiring locally usually means hiring excellent AR engineers who expect you to supply the 3D model, which for a project that has drawings and no model is the expensive half of the job.

San Francisco is an unusual case in this category, and worth treating differently from every other city.

In most markets the honest finding is that no local augmented reality industry exists and the buyer is choosing between national providers regardless of the map. That is not true here. The Bay Area has real augmented reality companies with real engineering depth, and a developer searching locally will find them.

The catch is what they do. Almost all of that capability is aimed at consumer brands, retail, entertainment and enterprise. Very little of it is aimed at buildings that have not been built.

So the local question here is not whether talent exists. It is whether the talent that exists has ever had to place a proposed structure on a graded parcel with correct survey position and terrain, and answer to somebody who reads elevations.

Why this city is a strong case for augmented reality anyway

Set the vendor question aside for a moment, because the format itself fits San Francisco better than it fits most places.

The people whose opinion determines a project outcome here are local and physically present. That is the exact condition augmented reality requires and the only one under which it outperforms a rendering. In a market where the decisive audience is out of state investors, augmented reality is close to worthless. In a market where the decisive audience walks past the site, it is close to ideal.

The recurring objection is also the one augmented reality answers best. Height and mass relative to what is already there is a question about scale, and scale is what people read worst from a screen and best from their own feet.

Three site conditions that break a generic AR build

Grade

This city is not flat and the software assumes it is. Phone augmented reality estimates a ground plane from the camera feed, and on a sloped street that estimate degrades quickly. A building placed against a wrong ground plane reads as floating or half buried, and on a steep block the error can approach a storey, which is precisely the dimension under dispute.

Correct terrain in the model, derived from survey rather than assumed, is not an optional refinement here. It is the difference between a persuasive experience and one that undermines the project.

Tight streets and short viewing distances

On a narrow street there is nowhere to stand far enough back to see a whole building. The viewer looks up, the phone tilts, and tracking quality falls exactly when the view gets interesting.

The practical answer is to plan the viewing positions in advance, choose the two or three places where the experience actually works, and mark them. An augmented reality preview with a designed route beats one that assumes the viewer will find the right spot.

Light and fog

Materials tuned under a rendered sky can look wrong under a real one, and this city supplies an unusually wide range of real ones. Glass in particular behaves differently in flat marine layer than in direct afternoon sun, and a facade that reads as elegant in a rendering can read as black or as mirror on site.

How this list was put together

Entries were identified through public research and are reachable products, platforms or services. Local augmented reality agencies, national platforms, surveyed field systems and done for you production appear together because a Bay Area project team compares them as one decision even though they solve different halves of the problem.

Criterion	What we looked for
Local presence	Whether the company has a genuine Bay Area presence or serves the market remotely.
Sector experience	Brand and enterprise augmented reality, or property and building work.
Who builds the model	Whether the quote assumes you supply usable geometry.
Terrain and position	Whether slope and surveyed placement are handled or assumed away.
Stated limits	Where each option stops being the right answer.

Editorial note: Rendimension publishes this guide and appears on it. We work remotely on Bay Area projects and say so plainly rather than implying a local office. We place a platform first because it is not a competing purchase, and we list ourselves in the narrow niche we serve rather than at the top. Every other entry is an independent company we do not control.

1. inCitu

First because it is a platform rather than a production studio, so it is not competing for the same scope, and because its use case maps closely onto how development decisions actually get made in this city.

The platform converts planning data into a full scale augmented reality layer that anybody can see by scanning a QR code while standing at the site. Published work spans more than four thousand construction sites, including municipal collaborations, with deployments across New York, California and beyond.

The reason it fits San Francisco specifically is that the audience which decides the fate of a project here is overwhelmingly local and physically present. Neighbours, merchant associations and hearing attendees are not looking at a deck in another state. They live on the block, and a full scale view from the sidewalk answers the height question in a way no elevation drawing does.

It works from data that already exists in a readable form, so a project without documented geometry has a modelling problem to solve before any platform helps.

2. Rendimension

Second, in the niche we serve: done for you augmented reality previews of buildings that do not exist yet, built from the drawings, delivered to teams working on Bay Area projects.

We work remotely and say so directly, because the honest structure of this market rewards it. The augmented reality talent in San Francisco is genuinely the deepest in the country and it is pointed at consumer brands, retail and enterprise rather than at real estate development. A developer hiring a local AR agency is usually hiring excellent AR engineers who will need the project explained to them, and who will expect the 3D model to arrive from somewhere else.

That is the gap we fill. We start from construction documents rather than from a finished model, which is the state most projects are actually in.

For Bay Area work specifically, the model has to carry the ground. Flat site assumptions fall apart on this topography, and a building placed without correct terrain reads as floating or sunken the moment a viewer walks half a block.

Declared terms rather than claims: first visuals in 48 to 72 hours, and reasonable revisions are included at no extra charge. We do not raise capital, we do not obtain approvals and we do not guarantee them, and we do not sell or lease property.

3. Rock Paper Reality

An augmented reality design agency working with startups and large brands, and one of the reasons San Francisco has a real local AR market rather than a nominal one.

The capability here is genuine and deep. Agencies of this type ship augmented reality that works reliably on consumer hardware at scale, which is harder than it sounds and is a different discipline from making something look good in a demo.

The mismatch for a developer is domain rather than skill. A retail AR campaign and a proposed building on a hillside parcel share a technology stack and share almost nothing else in terms of what has to be correct: terrain, survey position, glazing behaviour in real sky, and a model derived from a document set.

Worth approaching with a clear scope that states who supplies the model, because the assumption in brand work is usually that the client does.

4. Treeview

An AR, VR and XR studio building immersive work for enterprise customers, representative of the studio route in this market.

The studio route buys capability rather than a product, which suits a project that needs something no existing platform does. It also means the deliverable is defined by the brief, so a vague brief produces an expensive surprise.

The useful question for any general immersive studio is how many real estate projects they have completed, because the technical work transfers between sectors and the domain knowledge does not.

5. CitrusBits

A development company offering augmented reality and XR application work alongside mobile development, which is the vendor profile a Bay Area developer will encounter most frequently.

This is application development. If what the project needs is a branded app a sales team can hand to visitors, that is exactly the right kind of vendor to talk to.

If what the project needs is a convincing building, the application is the smaller half of the work and the model is the larger half. Two quotes that differ wildly almost always differ on that split rather than on quality.

6. Transparent House

A visualization studio producing architectural rendering and interactive work, included because the San Francisco visualization market is mature and augmented reality is only one output of it.

The relevant point for a project team is sequencing. The model is the durable asset. Once it exists, it produces stills, animation, a tour and an augmented reality preview, and the marginal cost of the fourth output is far lower than the first.

A team that commissions augmented reality in isolation, from a vendor who will not release the model, has bought the most perishable output first and kept none of the value.

7. Trimble SiteVision

The correct purchase when position has to be surveyed rather than approximate, which on this terrain matters more than it does almost anywhere else.

SiteVision pairs high accuracy GNSS with a handheld phone rig so the model sits by coordinate rather than by visual guess, and the 2026.10 release fuses lidar scan data into the same Trimble Connect workflow.

The San Francisco specific reason to care is slope. Phone augmented reality estimates the ground plane from what the camera sees, and a camera looking down a steep street is being asked a genuinely hard question. On a level lot the error is cosmetic. On a grade it can put a building a full storey out, which is the exact dimension everybody in the room is arguing about.

Local or remote: what actually changes

Worth being concrete, because proximity is oversold in this category and occasionally undersold.

What genuinely benefits from local presence: capturing the site, verifying sightlines from real positions, and being physically present when a sales team or a community group uses the experience for the first time.

What does not: modelling, optimisation, material work, lighting and delivery. That work happens on a screen and the address of the screen is irrelevant.

The reasonable arrangement for most Bay Area projects is remote production with someone local for the site day, which is how a large share of visualization work is already procured here.

What to ask a San Francisco AR vendor

How many building projects have you delivered, as opposed to brand or retail experiences. The answer changes the risk profile more than any other question.

How do you handle terrain, and do you work from survey data or estimate the ground plane. On this topography an evasive answer is disqualifying.

Who builds the model if we only have drawings. The default assumption in brand augmented reality is that the client supplies the asset, and that assumption quietly excludes the largest cost.

What happens when the tooling changes. Adobe Aero, a mainstream augmented reality authoring tool from a major vendor, reached end of support on 6 November 2025, and anything built to depend on one app inherits that risk.

When to spend the money somewhere else

If the buyers are out of state, augmented reality is the wrong purchase and a 360 tour or a set of renderings does far more for the same money.

If the project has no site access, there is nowhere to stand and the format has nothing to work with.

If only one visual asset is affordable, it should be a rendering, because a rendering reaches everybody and an augmented reality preview reaches whoever visits.

Frequently asked questions

Are there real augmented reality companies in San Francisco?

Yes, and unusually so. The Bay Area has genuine augmented reality agencies and studios with deep engineering capability. Most of that capability serves consumer brands, retail and enterprise rather than property development, so the question is sector fit rather than whether local talent exists.

Does the hilly terrain affect augmented reality?

Significantly. Phone augmented reality estimates a ground plane from the camera feed and that estimate degrades on slope. On a steep block the placement error can approach a storey, which is the exact dimension people argue about, so terrain from survey data rather than assumption is essential here.

Should we hire locally or remotely?

Modelling, optimisation, lighting and delivery are screen work and location neutral. Site capture, sightline verification and being present the first time a sales team uses the experience benefit from someone local. Remote production with a local site day is the common arrangement.

Is augmented reality useful for community meetings here?

It is one of the strongest uses of the format, because the recurring objection is height and mass relative to existing context and the audience is local by definition. A shared full scale view removes the argument about whose mental image is accurate.

What do we need before an AR preview can be produced?

A 3D model, or the drawings and survey information to build one from. Most projects have drawings rather than an augmented reality ready model, so realistic scope is model production plus mobile optimisation, not simply publishing an existing file.