Optional upgrade path to C1/C2. CDP screencast (JPEG over SSE) is fine for observation but low-fps/lossy under heavy motion. Commercial services (Steel) moved to WebRTC + H.264 @25fps for headful sessions.
Scope
- Evaluate WebRTC signaling (requires a server + ICE) vs the cost/dep weight for a CLI/MCP tool.
- Likely a heavy add (signaling, codecs) — assess whether it's worth it vs improving the SSE path (everyNthFrame tuning, WebP frames).
Acceptance
- Decision doc: WebRTC vs improved screencast, with a recommendation.
- Only implement if the fidelity gain justifies the dependency footprint.
Priority — low; nice-to-have.
Optional upgrade path to C1/C2. CDP screencast (JPEG over SSE) is fine for observation but low-fps/lossy under heavy motion. Commercial services (Steel) moved to WebRTC + H.264 @25fps for headful sessions.
Scope
Acceptance
Priority — low; nice-to-have.