📣AnnouncementProject Auriga · May 18, 2026

Project Auriga — Contributor Slots Filled

Thank you to everyone who applied. Project Auriga has reached full contributor capacity and applications are now closed. We will reopen slots as the program scales. Follow Flipside AI on Facebook to be notified the moment new slots become available.

Follow Flipside AI
for updates & slot openings
Resubmit your application

Were you asked to resubmit?

Applications are currently closed to new applicants. If received an email asking you to resubmit, enter your email below to access the form.

Active Programs

Project Registry

Each project runs independently with its own scope, unit of production, contributor criteria, and compensation structure. Client identities may be confidential — codenames are used in all contributor-facing materials.

Auriga
Full

Home environment data collection. Egocentric video of household chores and indoor manipulation — cleaning, cooking, object handling, domestic routines in real home settings.

Egocentric VideoPhone-BasedHome · IndoorQualification Required
Indoor · Scope TBD
Gaia
Preparing

3D scene capture for world model training. Orbital iPhone video of real-world environments — streets, markets, transit, parks, interiors. Two-pass capture protocol with gimbal. Static scenes, 60–90 sec per scene.

3D Scene CaptureiPhone · GimbalOutdoor + Indoor · WorldQualification Required
Orbital · iPhone + Gimbal
Nova
Premium

Premium egocentric data collection using iPhone ultra-wide video capture. Operator-collectors record natural, task-based activities across real-world professional environments using a qualifying iPhone 11 or newer. Global sign-ups open.

EGOCENTRIC · ULTRA-WIDE VIDEO12 NON-HOME TASK ENVIRONMENTSQUALIFICATION REQUIRED
iPhone 11 or newer · Head-mounted · 0.5× Ultra-wideSign Up Now
Vega
Internal

Nexscient Labs's proprietary training dataset. Mixed-environment egocentric collection — home chores and indoor tasks alongside outdoor and real-world activity. Broad scene coverage.

Egocentric VideoPhone-BasedIndoor + OutdoorQualification Required
Mixed · Scope TBD
Nova
EGOCENTRIC · ULTRA-WIDE VIDEO12 NON-HOME TASK ENVIRONMENTSQualification required

Nova is a premium egocentric video collection program designed for small and medium businesses where operators record their actual workplace operations. Participation requires an iPhone 11 or newer with a native 0.5× ultra-wide rear camera lens mounted on a head harness. All tasks must be completed in non-home, professional environments. Global sign-ups are open, and qualification is assessed via a structured onboarding review and an equipment verification process.

Rig specification — required hardware
Camera
iPhone 11 or newer · Rear 0.5× ultra-wide lens · Head-mounted, forward-facing · No external lenses or mods · iPhone SE and iPhone 16e not eligible
Mount
Head-mount harness with 1/4" adapter · Front-facing, egocentric POV · Landscape orientation
Computer
Laptop or PC (Windows or Mac) · Recommended for file offload and backup. GPS metadata is embedded by the iPhone at record time — enable Location Services for Camera before recording.
Storage
Portable SSD recommended (e.g. Samsung T7 Shield) · Min 1 TB · Used as intermediate offload destination before Portal upload
Power
65 W USB-C portable power bank · Min 25,000 mAh
Capture specification
Resolution
≥ 1080p RGB (1080p at 30 FPS minimum · 4K preferred where storage allows)
Frame Rate
≥ 30 FPS · 60 FPS preferred
Lens
0.5× ultra-wide (rear camera) · Single lens only · Spatial Video and stereo modes not permitted
GPS / Location
Location Services must be ON for the Camera app · Precise Location enabled · GPS is embedded at record time by the iPhone — videos recorded without GPS enabled will be rejected on upload
Episode Length
Short clips 30 s – 5 min · Long clips up to 30 min · No concatenated unrelated tasks
Format
MOV · H.264 (Most Compatible) or HEVC/H.265 (High Efficiency) · Landscape orientation · Audio off unless assignment specifies
Task diversity — required environment coverage
Operator-collectors must be able to perform natural, task-based activities across a minimum of 4 distinct task categories from the list below. Operators covering more categories are prioritised. No single category may exceed 20% of an operator's submitted volume.
Cleaning & SanitationFood & Beverage PrepRetail & Consumer GoodsPacking & LogisticsBeauty & Personal CareConstruction & RepairHealthcare & PharmacyChildcare & EducationLight AssemblyOffice & Admin TasksOutdoor Real-World
Operator requirementsOperator-collectors must be 18 or older, able to perform continuous task-based activity for 30+ minutes while wearing the head-mount rig, and have access to at least one qualifying professional environment (workplace, retail, or similar real-world setting). A laptop or PC is required for USB file transfer to preserve GPS metadata. Prior experience with data collection equipment is an advantage but not required — full rig training can be provided. Global sign-ups welcome.
Gaia
iPhone · GimbalOrbital CaptureQualification Spec TBD

Gaia uses an orbital 3D scene capture protocol — not head-mounted egocentric recording. Contributors use an iPhone (14 or newer) on a gimbal to record two-pass orbits around static real-world scenes. The qualification pathway will be published when the project scope is confirmed.

Device
iPhone 14 or newer
Equipment
Gimbal recommended
Orientation
Landscape
Duration
60–90 sec per scene
Protocol
Two-pass orbital (overhead + waist)
Scene
Static · 5+ distinct 3D objects
Vega · Auriga
Phone-basedQualification via link

Applicants for Vega and Auriga submit a link to a short qualification clip. Automatically checked for technical criteria, then human-reviewed for mount, hands, and content. Personal details collected only after this step passes.

Duration
Under 15 seconds
Resolution
1080p minimum
Frame Rate
30fps minimum
Orientation
Landscape
Zoom
0.5x or wider
Mount
Head-mounted, egocentric
Content
Both hands moving object across surface
Submission
Link — Google Drive, iCloud, Dropbox, OneDrive
Set link to "anyone with link can view." Keep it accessible until review is complete.
Auto checks: resolution · fps · orientation. Human review: mount · hands · object movement. One retry on fail.