WebView evaluation: Stay with KMP, implement CEF mitigation for Desktop

[Copilot]

Context

PR #36 revealed KMP's Desktop WebView (JavaFX WebKit) inadequate for modern SPAs like GitHub Copilot Agents. Evaluated whether to pivot frameworks or stay with KMP.

Decision

Stay with KMP + implement JCEF mitigation

Android WebView (Chromium): production-ready ⭐⭐⭐⭐⭐
Desktop WebView (JavaFX WebKit): limited ⭐⭐☆☆☆

Issue is localized (one platform, one component). Pivot costs months; mitigation takes 3-5 days.

Deliverables

docs/WEBVIEW_EVALUATION.md (691 lines)

Comprehensive analysis:

• KMP WebView bifurcation (Android excellent, Desktop limited)
• 11 frameworks evaluated (Qt, React Native, Flutter, Electron, .NET MAUI, etc.)
• Comparative matrix with WebView quality ratings
• 4 mitigation options with effort estimates
• Decision factors with weighted scoring

Key finding: Qt has best WebView (QtWebEngine everywhere), but poor mobile ecosystem and wearable support. No framework is perfect—KMP's strengths (WebRTC, audio, wearables) justify staying.

docs/WEBVIEW_DECISION_SUMMARY.md (176 lines)

Executive summary with acceptance criteria status, rationale, next steps.

Updated docs/FRAMEWORK_EVALUATION.md

Added WebView quality column to comparison matrices. 8 cross-references to evaluation.

Updated README.md

Linked to evaluation documents.

Recommended Mitigation

Phase 1: JCEF (Java Chromium Embedded Framework) for Desktop

• Replace JavaFX WebView with Chromium-based JCEF
• Proven technology (IntelliJ IDEA uses it)
• Effort: 3-5 days

Phase 2 (Fallback): External browser for Desktop Agent Tab

• Simple "Open in Browser" button
• Effort: 1 day

Framework Comparison

Framework	WebView	Trade-off
KMP	★★★☆☆ (split)	Mitigatable, keeps native strengths
Qt	⭐⭐⭐⭐⭐	Poor mobile/wearable ecosystem
React Native	⭐⭐⭐⭐☆	Desktop needs separate Electron app
Flutter	★★★☆☆	Same Desktop WebView issues
Electron	⭐⭐⭐⭐⭐	No mobile (blocker)
.NET MAUI	⭐⭐⭐⭐☆	WebKit on macOS/iOS

Mobile-first use case prioritizes Android (where KMP WebView excels). Agent Tab is secondary feature; core voice functionality unaffected.

Original prompt

---

This section details on the original issue you should resolve

<issue_title>Re-evaluate WebView solution for Codeoba: Is Kotlin Multi Platform's WebView viable, or should we pivot to Qt/other frameworks?</issue_title>
<issue_description>## Summary
Recent implementation work reveals that Kotlin Multi Platform’s (KMP) WebView—especially on Desktop (which uses a Java-based browser component)—is of inferior quality for our primary use case: embedding https://github.com/copilot/agents and monitoring/interacting with CoPilot Agent progress, then reporting notifications to the user. This is critical to our real-time agent UI/notification requirements. There are performance, compatibility, and extensibility issues compared to alternatives.

Context

• On 2025-12-11, we decided to move forward with Kotlin Multi Platform (KMP) for Codeoba, largely based on PR Add comprehensive cross-platform framework evaluation with Codeoba-specific requirements #3 and the original framework evaluation.
• 12 days post-decision, major limitations with KMP’s embedded webview became evident: the Java browser component does not work well with modern/complex websites like copilot/agents.
• A robust, extensible embedded web browser is key for Codeoba, for real-time interaction, programmatic DOM inspection & notification handling.

Questions to Revisit

• Is KMP’s WebView sufficient for our needs, or is it a blocker for full-featured agent UI/notification integration?
• Is Qt (with its modern QtWebEngine) a better fit for this requirement, or does it have significant tradeoffs?
• How do other cross-platform frameworks considered in PR Add comprehensive cross-platform framework evaluation with Codeoba-specific requirements #3 (see docs/FRAMEWORK_EVALUATION.md) fare in this area specifically?
  • Flutter
  • React Native
  • Electron
  • .NET MAUI
  • Others?
• What is the recommended path forward if KMP’s WebView is a non-starter?

Acceptance Criteria

• Documented review of KMP’s current WebView support and limitations for our use case
• Comparative matrix of embedded webview/browser capability for all major candidate frameworks
• Clear recommendation (stay or pivot), with rationale
• Linked supporting evidence/discussion
• Proposed mitigation plan if pivot required

Related Documentation

• PR Add comprehensive cross-platform framework evaluation with Codeoba-specific requirements #3: Add comprehensive cross-platform framework evaluation with Codeoba-specific requirements #3
• docs/FRAMEWORK_EVALUATION.md

---

This issue replaces the previously settled framework decision, in light of new technical blockers. Immediate investigation and discussion required.</issue_description>

Comments on the Issue (you are @copilot in this section)

• Fixes Re-evaluate WebView solution for Codeoba: Is Kotlin Multi Platform's WebView viable, or should we pivot to Qt/other frameworks? #38

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.

[Copilot]

Pull request overview

This PR provides a comprehensive evaluation of WebView technology across cross-platform frameworks in response to Desktop WebView limitations discovered during PR #36. The analysis concludes that Codeoba should stay with Kotlin Multiplatform (KMP) and implement mitigation strategies rather than pivot to a different framework.

Key changes:

• Added comprehensive WebView evaluation documenting KMP's bifurcated WebView support (excellent on Android, limited on Desktop)
• Evaluated 11 alternative frameworks with detailed WebView capability comparison
• Recommended JCEF (Java Chromium Embedded Framework) as primary mitigation strategy for Desktop

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 3 comments.

File	Description
docs/WEBVIEW_EVALUATION.md	Comprehensive 691-line analysis of WebView capabilities across frameworks, including technical assessment, comparative matrices, mitigation strategies, and decision rationale
docs/WEBVIEW_DECISION_SUMMARY.md	Executive summary (176 lines) documenting acceptance criteria status, key findings, and recommended next steps for stakeholder review
docs/FRAMEWORK_EVALUATION.md	Updated existing framework evaluation with WebView quality ratings column and cross-references to the new WebView evaluation
README.md	Added links to the new WebView evaluation documentation in the documentation section

[Copilot]

The issue number placeholder "XX" should be replaced with the actual issue number. Based on the PR description, this should be issue #38.

Suggested change

	**Issue:** [Re-evaluate WebView solution for Codeoba](https://github.com/LookAtWhatAiCanDo/Codeoba/issues/XX)
	**Issue:** [Re-evaluate WebView solution for Codeoba](https://github.com/LookAtWhatAiCanDo/Codeoba/issues/38)

[Copilot]

The issue number placeholder "XX" should be replaced with the actual issue number. Based on the PR description, this should be issue #38.

[Copilot]

Inconsistent line count stated. The summary mentions "691 lines" but the document states "1300+ lines" in the summary text. The actual line count is 691, so the reference on line 131 should say "691 lines" instead of "1300+ lines".

Suggested change

	- **[WEBVIEW_EVALUATION.md](WEBVIEW_EVALUATION.md)** - Comprehensive evaluation (1300+ lines)
	- **[WEBVIEW_EVALUATION.md](WEBVIEW_EVALUATION.md)** - Comprehensive evaluation (691 lines)