Free, open-source screen sharing for private live streams with friends. Watch together in a browser or self-host Piik. 免费开源的私密屏幕共享,支持游戏直播、一起看电影与自部署。
Verification Status
Last updated: 2026-10-26
This file summarizes evidence boundaries that materially affect product claims. Status owns candidate and deployment readiness; research owns measurements and platform details.
Current Evidence
Windows App and Browser are the primary acceptance targets. Room voice passed the normal Host/Viewer flows on P2P and embedded SFU, including channel activation, forwarding, source replacement and recovery. Controlled voice checks cover Browser/App endpoint combinations, SFU fallback and UDP interruption; they do not establish a population connection rate. Native H264/VP8, constrained-network recovery, the media lifecycle, responsive UI and packaging checks also passed. Exact results stay with their measured artifacts and revision.
The owner accepted the earlier scoped interaction, Native and Browser media phases, including NAT and Auto codec behavior.
Device Evidence Boundary
- macOS/Linux builds and bounded native checks are distinct from physical capture and input. Physical platform coverage remains limited.
- Android's system-browser App passed emulator screen/camera sharing with concurrent room voice. Browser consent returns to the same tab; source replacement, cancellation and background return preserve the share. Local/public startup and launcher reentry also have emulator evidence. Full-App checks now cover concurrent original/lower encoding, live-budget reduction and recovery, static-source late joins, lower-output retirement, capture stop and restart. The capture assessment owns measured bitrate bounds and the stock FrameDropper's saturation limit. Phone/tablet/TV hardware coverage and physical audio remain limited.
- The separate iOS system-browser implementation passes Apple SDK device and simulator builds and its session/encoder checks. Physical capture, background lifetime and hosted-site linkage remain in the held iOS scope.
- Windows HDR-to-SDR has synthetic GPU and decoded-video evidence; it does not establish Browser-capture HDR conversion or end-to-end HDR. The fidelity assessment owns that boundary.
- Audio-exclusion availability follows actual platform capability. The audio assessment owns supported interfaces and isolation evidence.
- Room-voice acceptance uses synthetic microphones. It establishes the route and membership lifecycle, not acoustic echo performance on every physical device.
- Remote-control UI and routed input have local evidence. OS permissions, protected/admin applications and game compatibility retain the limits in the input assessment.
- A two-device Windows virtual-LAN trial established direct packaged-App web access and remote video after scoped firewall permission. Other networks must likewise carry HTTP/WebSocket and media UDP. Virtual LAN research owns the measured setup.
Adaptation Boundaries
WebRTC and LiveKit own adaptation. The encoder-pool research and network controls distinguish Native recovery fixes from Browser-controlled recovery and sustained-loss limits. Browser scripts cannot tune the encoder's internal QP thresholds.
Interpretation
Evidence applies to its tested revision, device and workload. Builds and local controls establish their named behavior; they do not promise every remote network or device or a population NAT success rate. Browser/OS suspension is not a keepalive guarantee.