> ## Documentation Index
> Fetch the complete documentation index at: https://dev-doc.dualmindlab.tech/llms.txt
> Use this file to discover all available pages before exploring further.

# Roadmap

> Planned features and improvements for DualMind Lab — the blind AI model comparison arena.

# Product Roadmap

A look at what's coming next for DualMind Lab. Features are organized by priority and estimated timeline.

<Info>
  This roadmap is subject to change based on community feedback and development priorities. Have a feature request? [Open an issue on GitHub](https://github.com/HarshBhanushali07).
</Info>

## In progress

<CardGroup cols={2}>
  <Card title="UI Revamp" icon="paintbrush">
    **Priority**: High

    Complete redesign of the user interface with modern aesthetics, improved accessibility, and better mobile responsiveness.

    * New design system with consistent components
    * Dark/light mode improvements
    * Mobile-first responsive layout
    * Status: \~80% complete, launching soon
  </Card>

  <Card title="ELO Rating System" icon="chart-line">
    **Priority**: High

    Implement a proper <Tooltip tip="Elo rating system adapted from chess for ranking AI models based on pairwise comparisons">ELO rating</Tooltip> algorithm for model rankings instead of simple win-rate percentages.

    * Bayesian ELO with confidence intervals
    * Rating decay for inactive models
    * Historical rating charts
  </Card>
</CardGroup>

## Planned — Q2 2026

<Steps>
  <Step title="Additional AI Providers">
    Integrate OpenAI, Anthropic, Google Gemini, and Mistral as first-class providers alongside Groq and Bytez.

    | Provider  | Models              | Status  |
    | --------- | ------------------- | ------- |
    | Groq      | Llama, Mixtral      | Live    |
    | Bytez     | Various             | Live    |
    | OpenAI    | GPT-4o, GPT-4o-mini | Planned |
    | Anthropic | Claude 3.5 Sonnet   | Planned |
    | Google    | Gemini 2.0          | Planned |
  </Step>

  <Step title="Public API Keys">
    Allow developers to generate API keys for programmatic access without requiring Supabase Auth session tokens.

    * API key management dashboard
    * Rate limiting per key
    * Usage analytics
  </Step>

  <Step title="Batch Comparisons">
    Run automated batch comparisons across model pairs with predefined prompt sets for systematic evaluation.
  </Step>

  <Step title="Export & Analytics">
    Export comparison data, vote history, and model statistics in CSV/JSON formats for external analysis.
  </Step>
</Steps>

## Planned — Q3 2026

<Steps>
  <Step title="Custom Evaluation Criteria">
    Define custom rubrics for model evaluation beyond simple preference voting — accuracy, creativity, helpfulness, safety.
  </Step>

  <Step title="Team Workspaces">
    Collaborative workspaces where teams can run comparisons, share results, and build consensus on model selection.
  </Step>

  <Step title="Webhook Integrations">
    Real-time notifications via webhooks when comparisons complete, models are added, or rankings change.
  </Step>

  <Step title="Plugin System">
    Extensible plugin architecture for custom providers, evaluation metrics, and UI components.
  </Step>
</Steps>

## Completed

<AccordionGroup>
  <Accordion title="v2.0 — Blind Comparison Arena">
    * Arena battle mode with anonymous model responses
    * Random, topper, and manual model selection
    * Real-time SSE streaming
    * Thread management with public sharing
    * Admin dashboard with full CRUD
    * Supabase Auth integration
  </Accordion>

  <Accordion title="v1.5 — Groq Provider Support">
    * Added Groq as primary AI provider
    * Llama 3.3 70B and Mixtral 8x7B models
    * Model status monitoring
    * Provider API key rotation
  </Accordion>

  <Accordion title="v1.0 — Initial Release">
    * Single chat mode
    * Basic thread management
    * User authentication
    * PostgreSQL database schema
  </Accordion>
</AccordionGroup>

## Request a feature

<Tip>
  We prioritize features based on community demand. If you'd like to see something on this roadmap, open a GitHub issue or contribute directly.
</Tip>

<CardGroup cols={2}>
  <Card title="Feature Request" icon="lightbulb" href="https://github.com/HarshBhanushali07">
    Suggest a new feature on GitHub
  </Card>

  <Card title="Contributing" icon="git-pull-request" href="/development/contributing">
    Contribute code directly
  </Card>
</CardGroup>