# LlmBench Roadmap and Development Plans

> See what LlmBench has already shipped and the current next steps for provider verification, prompt tooling, UX, widgets, and security.

LlmBench is an early prototype evolving quickly. This page follows the released state on the repository's default branch rather than older design plans.

## Completed

<CardGroup>
  <Card title="WebView runtime" icon="check">
    Hardened provider WebViews, added a small LRU pool, pauses inactive views, and evicts them under memory pressure while provider-owned session state persists.
  </Card>

  <Card title="KMP shared core" icon="check">
    Moved portable provider, model, profile, and prompt logic into the Kotlin Multiplatform `shared` module.
  </Card>

  <Card title="Uploads and sharing" icon="check">
    Added provider file uploads through Android's file picker plus share-sheet routing for text, images, and application files.
  </Card>

  <Card title="Native provider gateways" icon="check">
    Added direct Gemini, OpenAI, and Claude transports plus OpenRouter Free and AIHubMix gateway adapters.
  </Card>
  <Card title="Provider integration layer" icon="check">
    Added provider-scoped tweaks/userscripts, privacy-safe diagnostics, and generating/unread tracking for the first supported web providers.
  </Card>

  <Card title="Provider compatibility docs" icon="check">
    Documented full and partial WebView support, including cases where embedded sign-in or uploads still require verification.
  </Card>

  <Card title="CI and repository cleanup" icon="check">
    Normalized the app identity, removed generated/build-machine files, and added build/lint checks.
  </Card>
</CardGroup>

## Planned

<Steps>
  <Step title="Verify newer WebView providers">
    Finish embedded sign-in, upload-flow, and activity-probe verification for Qwen, Microsoft Copilot, Z.ai, Grok, Character.AI, Venice, and Meta AI.
  </Step>
  <Step title="Prompt and skill workflows">
    Add a local Markdown prompt vault plus editable/importable `SKILL.md` assets with safe capability gating.
  </Step>
  <Step title="First-party document and code tools">
    Bring selected Docbench and Codebench validation, repair, tokenizer, and hidden-text inspection features into compatible prompt/file workflows.
  </Step>
  <Step title="Modernize chat and navigation UX">
    Continue Material 3 Adaptive and Compose performance work, including large-screen list-detail layouts.
  </Step>
  <Step title="Widgets and notification replies">
    Add provider-aware chat/message widgets and notification replies where the transport can safely send in the background.
  </Step>
  <Step title="Security and privacy release gates">
    Keep sensitive storage, backup rules, WebView isolation, tool permissions, redacted previews, safe exports, and regression tests as cross-cutting release requirements.
  </Step>
</Steps>

## How to contribute

Useful contributions include provider verification, shared KMP improvements, adapter maintenance, prompt/skill tooling, security testing, and measured UI performance work.

<Tip>
  Check the repository's issue tracker and pull requests for the freshest implementation work before starting a larger change.
</Tip>