HOSHIMI CAMERA

Take photos with your virtual characters

Take photos with your virtual characters

HOSHIMI CAMERA blends your live camera feed with VRM avatars. Place a character in the shot, give her a pose, an action or an expression, and shoot. What lands in your gallery looks like she was standing there.

A phone showing the HOSHIMI CAMERA viewfinder with an anime character standing in a real room
Point, place, pose, shoot — the composite happens on the phone.

Everything stays on your phone — no account, no upload, no analytics SDK.

WHY IT FEELS DIFFERENT

Invite a character into your shot

Pick a character from your library or import your own VRM file. Drag, scale and rotate her until the framing feels right.

Pose, perform, express

Choose from 20+ expressions, saved poses and animated actions. Freeze any frame mid-motion and shoot that exact moment.

A studio in your pocket

Framing, beauty, denoise, filters, zoom, grid, flash and front/back camera — all in one panel, all live on the preview.

Your room is the stage

Switch between characters mid-scene and keep everyone in frame. The app composites the characters on top of the live camera feed, so the light and the mess of your room stay in the shot.

CORE CAPABILITIES

/ 01 — PLACE

Invite a character into your shot

Pick a character from your library or import your own VRM file, then place her in the frame by hand.

  • · Built-in cast, or import your own .vrm file
  • · Drag, scale and rotate until the framing feels right
  • · Switch characters without leaving the viewfinder
  • · Front and back camera, switchable at any time
Character panel: size and angle sliders, expression chips, action transport and pose buttons
Size, angle, expressions, actions and poses — all dialled in from one panel.

/ 02 — PERFORM

Pose, perform, express

Give her a pose, an action or an expression — and freeze the exact frame you want to shoot.

  • · 20+ expressions, saved poses and animated actions
  • · Scrub an action and stop on the frame you like
  • · Expressions and actions come along with VRM 0.x / 1.0 files
  • · Other glTF/GLB models import too, but usually arrive static
Viewfinder: a virtual character standing in a real room, with the camera HUD below
The viewfinder: your room, your light, and her standing in it.

/ 03 — CAMERA

A studio in your pocket

Camera controls that work live on the preview, not afterwards.

  • · Framing, beauty, denoise and filters
  • · Zoom, grid, flash, front/back camera
  • · Photo at three quality levels, video, timer and time-lapse
  • · Everything is applied to the live preview before you shoot
Camera panel: framing, beauty, filter and zoom controls over the live preview
Framing, beauty, filter and zoom — live, over the preview.

/ 04 — RESULT

Your room is the stage

The composite happens on the spot, and the finished photo lands straight in your gallery.

  • · Composited on the device at capture time
  • · Saved to Pictures/MagicCamera in your gallery
  • · Stills, clips and time-lapse
  • · No cloud, no account, nothing to sync
A finished photo: the same character composited into the camera shot
The finished photo — composited on the spot, nothing sent anywhere.

/ 05 — LIBRARY

Your cast, in one library

The built-in cast and everything you imported, with their actions and poses alongside.

  • · Character cards for the built-in cast
  • · Actions, poses and expressions listed per character
  • · Import .vrm / .vrma / .fbx / .json from creator platforms
  • · Available in 8 languages, following your system language on first launch
Library tab: character cards for the built-in cast, with actions, poses and expressions alongside
Library: each character with her actions, poses and expressions.

SPOTLIGHT

Who it's for

You built an avatar and you only ever see her inside a mirror of yourself. This puts her in real light, in a real room, and takes the photo.

  • — VTuber & VRoid creators
  • — Cosplay & photo fans — no green screen, no rig
  • — Creators in a hurry: stills and clips in a couple of taps
  • — Anyone who wants their phone camera to do the compositing

HOW IT COMPARES

How it compares

Getting a virtual character into a real photo is usually a two-machine job: capture on one side, composite on the other. This does both on the phone.

 HOSHIMI CAMERADesktop VTuber toolsGreen-screen video editors
Where it runsAndroid phoneWindows PCAny editor, after the shoot
Live placementYes — in the viewfinderYes, on the desktopNo
VRM expressions / posesYes, per characterYes, with a rigNo
Extra gearJust the phonePC + tracking setupGreen screen
AccountNoneNoneVaries by editor
PriceFreeVariesVaries

Desktop VTuber tools

PC-based avatar performance and capture (VSeeFace, Warudo and friends) — powerful, but they live on a Windows machine.

Pros

  • + Full body and face tracking with real hardware
  • + Deep control over scenes and cameras
  • + Mature ecosystem and plugins

Cons

  • − Needs a Windows PC and a tracking setup
  • − Not a phone camera app — no shooting on location
  • − Compositing and export happen on the desktop

Green-screen video editors

Chroma-key compositing in post (CapCut and similar) — works with any footage, but the shoot comes first.

Pros

  • + Works with any footage, camera or avatar
  • + Full editing timeline afterwards
  • + Nothing to place during the shot

Cons

  • − You need a green screen and a separate shoot
  • − No live placement — you frame the character blind
  • − No VRM expressions, poses or actions to drive

If you want the character in a real place right now, shoot it on the phone. If you need tracked performance or a full post-production timeline, the desktop route is still the right tool.

NOTES FROM USE

Notes from using it

What we keep coming back to, and what to know before you start.

  1. Frame first, pose second

    Sep 30, 2026

    Placing the character before choosing a pose almost always looks better: you can see how the light in the room falls on her, and adjust the angle to match.

  2. Buying characters is a separate thing

    Sep 30, 2026

    Characters and actions from creator platforms are a deal between you and the creator. This app only imports them locally; what you may do with them afterwards follows their terms.

Related writing

We have not written about HOSHIMI CAMERA yet — the first notes will show up here.

System requirements

Operating system
Android 8.0 or later (API 26)
Architecture
arm64-v8a (64-bit)
Installer
APK — about 103.5 MB (108,545,098 bytes)
Version
1.0 (versionCode 1)
Package
work.aihow.magiccamera
Character files
VRM 0.x / 1.0 (expressions, poses, actions); other glTF/GLB import static
Photos
Saved to Pictures/MagicCamera in the device gallery
App language
English, 中文, 日本語, 한국어, Deutsch, Français, Español, Italiano
Licence
Free — no ads, no in-app purchases, no subscription

The APK is about 100 MB, so we suggest downloading over Wi-Fi. After downloading in a browser, your phone will ask you to allow installing apps from unknown sources for that file.

Free

HOSHIMI CAMERA is free: no ads, no in-app purchases and no subscription — every feature is in the download. The only things you might pay for are third-party characters or actions you buy on creator platforms such as BOOTH; that is between you and the creator, and this page handles no payment for it.

Pricing is unavailable right now — please try again in a moment.

Everything, actually

There is nothing to unlock. Every module, every character slot and every capture mode is in the free download.

  • · All capture modes: photo, video, timer, time-lapse
  • · All 20+ expressions, poses and actions
  • · Import your own .vrm characters
  • · 8 languages, no account
  • · No watermark, no trial countdown

QUESTIONS

FAQ

Is it really free?

Yes. The app is free, with no ads, no in-app purchases and no subscription — every feature is in the download. The only things you might pay for are third-party characters or actions you choose to buy on creator platforms such as BOOTH; that is a transaction between you and the creator, and this page does not handle any payment for it.

Do I need an account?

No. There is no sign-up, no login and nothing to sync.

Which character files can I use?

VRM 0.x and VRM 1.0 avatars import fully — expressions, poses and actions included. Other glTF/GLB models can be imported too, but they may come in static: those files often carry no facial blend shapes and no humanoid rig, so expressions and actions stay unavailable for them.

Where do my photos go?

Straight to your device's gallery, in the Pictures/MagicCamera folder, so they show up in whatever photo app you normally use.

Does it work with the front camera?

Yes — flip between the front and back camera at any time, and switch characters without leaving the viewfinder.

Which languages does the app speak?

English, 中文, 日本語, 한국어, Deutsch, Français, Español and Italiano. It picks up your system language on first launch and remembers your choice after that.

How do I add my own character or action?

Download the file (.vrm, .vrma, .fbx or .json) from a creator platform such as BOOTH, then use the import entry inside the app. What you may do with the file is decided by its creator — the app only imports it locally.

Download

Grab the APK, then open it on your phone. Version, size and SHA-256 below are for the file you are actually downloading.

The installer is not published yet. It will appear here as soon as the build is ready.