[Bug]: Screen recording fails on Niri/Wayland with PipeWire (no more input formats)

[talison-cardoso]

Search existing issues

• I have searched the existing issues

Describe the bug

Environment

• OpenScreen: 1.9.0-rc.3 (also reproducible on 1.8.0)

• Distribution: Arch Linux (CachyOS)

• Desktop: Niri

• Session: Wayland

• PipeWire: active

• xdg-desktop-portal: active

• Installed portal backends:

  • xdg-desktop-portal-gnome
  • xdg-desktop-portal-gtk

Description

OpenScreen is unable to start a screen recording.

The screen selection dialog appears normally, but after confirming the screen/window, the recording never starts and the application reports a PipeWire stream error.

The issue is reproducible on both 1.8.0 and 1.9.0-rc.3.

Logs

OpenScreen:

[native-linux] starting capture {
  outputPath: '/home/.../recording.mp4',
  cursor: { mode: 'editable-overlay' },
  audio: {
    system: { enabled: false },
    microphone: { enabled: false, gain: 1.4 }
  },
  video: { fps: 60 }
}

[capture-linux] stream-error:
PipeWire stream reported an error in state error: no more input formats

PipeWire:

pw.context:
  Rectangle 2560x1080

pw.context:
  Fraction 59978/1000

pw.link:
  negotiating -> error no more input formats

Additional information

• pipewire.service is running normally.
• xdg-desktop-portal.service is running normally.
• The portal logs do not report ScreenCast-related errors.
• Installed portal backends are gnome and gtk.
• The problem occurs before any recording starts.

Please let me know if you would like additional debug logs (for example, PIPEWIRE_DEBUG=5) or if there are other diagnostics I can collect.

Expected behavior

Note: The report above was prepared with the assistance of ChatGPT. After providing the application logs, environment details, and the troubleshooting steps I performed, I asked the AI to organize the information into a concise report suitable for opening this issue.

The technical information (logs, environment, and observed behavior) comes from my own testing. The AI was only used to help structure and present the information more clearly.

My expectation is that OpenScreen should work correctly in this environment (Arch Linux + Niri + Wayland), allowing screen recordings to start and complete successfully.

To Reproduce

No response

Screenshots

No response

OS

Linux

OS Version

No response

Other OS

No response

Browser

None

Browser Version

No response

Other Browser

No response

Device Type

Laptop

Other Device

No response

Additional context

No response