
FACE
Target gallery, multi-ref uploads, drag targets. Pipe live capture or file-based sources.
JINKUSU CAM
INIT::signal_override
JINKUSU CAM replaces what your OS sees: swapped faces on webcam or file sources, OBS Virtual Camera out, voice morphing, and mobile relay — gated by license key + TOTP.
Module map
Captured from the running client — six primary tabs, one control surface.

Target gallery, multi-ref uploads, drag targets. Pipe live capture or file-based sources.

Resolution tiers, swap quality, background kill/enhance — OBS Virtual Camera as the exit node.

Voice presets, pitch shift, denoise chain — meters show live levels in the sidebar.

Phone-side relay: push virtual loops over ADB/Wi‑Fi from the desktop ANDROID module.

BlueStacks / LDPlayer ingress — OBS VCam becomes the camera inside emulated apps.

Still-image animation via landmarks — export GIF or MP4 without a live source.
Core stack
InsightFace pipeline on live or file input — Fast through Ultra quality tiers.
Standard virtual device any downstream app can bind — calls, browsers, emulators.
On-device morph + denoise presets. Headphones recommended to kill feedback.
Background blur/replace, expression track, frame enhance — one graph.
License key + authenticator. Seats, sessions, and builds managed centrally.
Package com.jinkusu.cam.app — emulator path via OBS VCam.
Boot sequence
OBS Studio, VC++ x64, current GPU drivers. FFmpeg optional for file sources.
Grab the Windows build from Access gate with license + TOTP.
FACE module: pick gallery faces, bind webcam or MP4/still input.
Hit START — confirm VCam: ON. Downstream apps select OBS Virtual Camera.
Host specs
Run locally on your hardware, or skip the GPU requirement with a rented RTX 4090 node.
| Component | Minimum | Recommended |
|---|---|---|
| OS | Windows 10 x64 | Windows 11 x64 |
| CPU | 4 cores · AVX2 | 8+ cores |
| RAM | 16 GB | 32 GB |
| GPU | DirectML (AMD / Intel) | NVIDIA RTX · CUDA |
| Storage | 8 GB free + models | SSD · 20 GB+ |
| OBS Studio | Required — Virtual Camera enabled | |
| Network | Stable connection for license validation | |
Mobile relay
Extends the Windows ML graph to a handset — loop files pushed from the ANDROID tab over ADB.
com.jinkusu.cam.app/sdcard/DCIM/Camera1/virtual.mp4
Plans · USD
Generate an order code, choose crypto, complete payment offline. 15% launch rate applied below.
Single seat
$500 $425 / month
Campaigns
$1,000 $850 / 3 months
Partners
$5,000 $4,250 one-time
Checkout opens Telegram @jinkususupport with your order code. Card payments are not handled on this site.
Cloud compute · add-on
No local CUDA or DirectML? Run JINKUSU CAM on a dedicated RTX 4090 · 32 GB VRAM node — billed monthly, provisioned after payment.
Software license is not included — pair with a Professional or bundle plan, or add vGPU to an existing seat via support.
Monthly rental
$150 / month
Order via Telegram @jinkususupport — same crypto checkout as software plans.
Access gate
Same credentials as the desktop binary: license key + TOTP.
Installer track: v8.0.1
> INIT::auth_gate
> AWAIT::credentials
No installer on file yet — check back after your operator publishes a build.
Payload index
Licensed binary via access gate. External deps for OBS, drivers, emulators, and models.
models/Get
PYPython 3.10+Developers onlyGet
Release track
Intel
Only through Access gate after license + TOTP auth. Public tabs list third-party deps.
Yes for Virtual Camera output — install OBS and enable its VCam before arming JINKUSU.
Key + authenticator code. Client validates against the API; SuperAdmin controls seats and builds.
No — select a plan, then send the pre-filled order to @jinkususupport on Telegram.
Yes — rent a dedicated Virtual GPU RTX 4090 (32 GB) for $150/month. You get remote access to a CUDA node tuned for the stack; software license sold separately.
Typically within 24 hours on business days once payment is confirmed on-chain. Message @jinkususupport with your order code for status.
Seat limits depend on your plan tier — configured by the operator when the key is issued. Contact support to add seats or transfer hardware.
Activated digital licenses are non-refundable. vGPU can be cancelled before the next monthly cycle. See Terms of Service.
License validation only — key, TOTP, device ID. Face/audio processing stays local. See Privacy Policy.
When published, it appears in the access gate after login.
Comms
Official channels for releases, licensing, and encrypted contact.
0502c796a13549c53da8727c14e900cb23703b49259357f6e3e215c542d0a75666
Encrypted messenger · paste ID in Session app
Pending order — resume checkout or dismiss.
Store this code — it ties your plan selection to the crypto payment step.
Opens chat with @jinkususupport — your order code, plan, and crypto choice are pre-filled.
Complete payment in Telegram with @jinkususupport. Your order code is saved locally for 7 days if you need to resume.