Skip to content

Latest commit

 

History

History
154 lines (107 loc) · 11.2 KB

README.md

File metadata and controls

154 lines (107 loc) · 11.2 KB

Alikel - Your Specialized AI Assistant Platform

Meet Leif & Max: Specialized AI assistants collaborating to enhance your productivity and fitness.
Live Demo: alikel.net

🚀 Overview

Alikel is an innovative AI platform featuring a suite of specialized assistants designed with distinct personalities and expertise. Unlike general-purpose AI, Alikel offers focused support through dedicated agents:

image image

Converse with them individually or together in a unique multi-agent experience, leveraging their combined strengths to tackle complex goals. Alikel seamlessly integrates conversational AI with practical tools, allowing for a dynamic and efficient way to manage your life.

✨ Key Features

Productivity Suite (Powered by Leif)

  • Intelligent Task Management: Seamlessly create, organize, and track tasks through conversation. Tasks generated by Leif automatically update your dedicated Tasks Page.
  • Integrated Calendar View: Visualize your schedule and upcoming tasks within the platform's calendar, providing a clear overview of your commitments.
  • Multiple Task Views: View tasks in list, grid, or calendar format to suit your workflow.
  • Action Tokens: Automate common productivity actions directly through conversation (e.g., setting timers, adding tasks).
  • Bulk Operations: Efficiently manage multiple tasks at once for planning and organization.
  • Time Management Tools: Integrated timers to help you stay focused and manage work sessions.

Fitness Coaching (Powered by Max)

  • Personalized Workout Plans: Tailored exercise routines based on your profile, goals, and available equipment.
  • Comprehensive Fitness Profiles: Track body measurements, weight, body fat percentage, and visualize progress over time.
  • Evidence-Based Guidance: Receive fitness and nutrition advice grounded in scientific research.
  • Motivational Support: Engage with an energetic coach designed to keep you on track.

Platform Capabilities

  • Multi-Agent Collaboration: Interact with Leif and Max simultaneously for integrated life planning.
  • Image Analysis: Upload images for contextual understanding and discussion
  • User Profiles: Personalize your interactions and receive tailored advice by creating user and fitness profiles.
  • Conversation History: Review past interactions and track ongoing dialogues.
  • Responsive Design: Access Alikel seamlessly across desktop, tablet, and mobile devices.

💡 Our Vision

We believe the future of AI lies in specialized agents that augment human capabilities, not replace them. Alikel embodies this vision by creating a collaborative ecosystem where distinct AI personalities work together, providing focused expertise within a unified, intuitive interface. Our goal is to build practical AI tools that automate the tedious while enhancing the human touch.

🤖 Technical Implementation

Alikel employs several techniques to achieve its specialized multi-agent functionality and seamless integration with UI tools:

  1. Multi-Agent Orchestration:

    • Distinct Personas via System Prompts: Each assistant (Leif, Max) operates with a detailed system prompt defining their unique personality, background, expertise, relationship with other agents, and conversational style (ai-chat.js, max-fitness-chat.js). This ensures specialized and consistent responses.
    • Contextual Awareness: User profile data (general and fitness-specific) is dynamically injected into the relevant assistant's prompt, allowing for personalized interactions without compromising the agent's core persona.
    • Shared History Simulation: Agents are made aware of their "shared history" and relationship through specific context provided in their prompts, enabling more natural and informed interactions between them.
    • Backend Conversation Management: Netlify Serverless Functions act as an orchestration layer. When multiple agents are active, the backend manages the turn-taking. For instance, the user message might be sent to Leif first. Leif's response is then added to the history, and this updated history (clearly marking Leif's contribution, e.g., [FROM LEIF]: ...) is sent to Max, allowing Max to react appropriately.
    • Role Differentiation: The conversation history sent to the LLM explicitly identifies the speaker (User, Leif, Max, System), allowing the active agent to understand the conversational context accurately.

  2. Action Token System (Leif):

    • Custom Token Format: Leif is instructed via his system prompt to output specific commands using a defined {{action:type:param1:param2...}} format when he intends to perform a productivity action (e.g., add task, start timer).
    • Client-Side Parsing: A dedicated React component (LeifActionParser.jsx) runs on the frontend. It scans Leif's responses for these action tokens.
    • Contextual Execution: Upon detecting a valid token, the parser extracts the action type and parameters. It then invokes corresponding functions within the shared ProductivityContext (e.g., suggestTask, startTimer).
    • User Confirmation & Feedback: For actions like adding tasks, a confirmation dialog (TaskConfirmationDialog.jsx) is presented to the user, allowing edits before finalizing. Once an action is processed (either directly or after confirmation), a system message is injected back into the chat to inform both the user and the AI agents that the action was completed.
    • Bulk Action Support: The parser handles both single actions and bulk actions (e.g., {{action:add_tasks:[{...},{...}]}}), enabling Leif to propose multiple tasks at once based on user requests.

  3. Integrated State Management:

    • React Context API: The ProductivityContext serves as a central hub for managing shared state related to tasks and timers. This allows different components (Chat Interface, Task Manager, Calendar View, Timers) to access and modify the same data seamlessly.
    • Local Persistence: User profiles, fitness data, conversations, and tasks are persisted in the browser's localStorage. This ensures data privacy (stored client-side) and allows users to resume their sessions.
    • Real-time UI Updates: When an action token is processed or a user interacts with the UI tools (e.g., completing a task on the Tasks Page), the context updates the state, triggering re-renders across relevant components for a synchronized experience.

  4. Serverless Backend:

    • API Abstraction: Netlify Functions provide a simple, scalable backend API layer. They handle communication with external AI services (Gemini, Moondream).
    • Secure Key Management: API keys for AI services are stored securely as environment variables in Netlify, never exposed on the frontend.
    • Specialized Endpoints: Dedicated functions (ai-chat.js, max-fitness-chat.js, moondream-analysis.js) handle requests for specific agents or functionalities, allowing for tailored logic and prompt injection.

This architecture allows for complex AI interactions and tool integration while maintaining distinct agent specializations and a responsive user experience.

🚀 Future Roadmap

We are constantly innovating. Key future enhancements include:

  • Google Calendar Integration: Allowing Leif to view your Google Calendar, understand your availability, suggest optimal times for tasks, and add events directly to your calendar (potentially via Google API or MCP Servers).
  • Expanding the Assistant Suite: Introducing new specialized assistants for different domains (e.g., culinary (Auguste!), career (Kei!)).
  • Advanced Analytics: Providing deeper insights into productivity patterns and fitness progress, similar to the graphs on DeepFit.
  • Fitness Action Tokens: Enabling Max to generate structured workout plans via action tokens, allowing users to save and manage them.
  • Feature Consolidation: Migrating functionality from DeepFit into the unified Alikel platform.
  • Data Portability: Implementing features to easily export and import all user data (conversations, profiles, tasks) to reinforce local storage privacy benefits.

🛠️ Technology Stack

Alikel is built with modern web technologies for a performant and engaging user experience:

  • Frontend: React 18 with Vite
  • Styling: Tailwind CSS with custom UI components
  • State Management: React Context API & Hooks
  • Animations: Framer Motion
  • Routing: React Router
  • Backend: Netlify Serverless Functions
  • AI Integration: Google Gemini API, Moondream Vision API

📸 Visuals

  1. Multi-Agent Chat Interface: Multi-Agent Chat

  2. Task Management (List/Grid View): Task List View

  3. Task Calendar View: Task Calendar View

  4. Fitness Profile Tracking: Fitness Profile Measurements Fitness Profile Form

💬 Usage Examples

Multi-Agent Collaboration: Multi-Agent Example 1 Multi-Agent Example 2 Multi-Agent Example 3

📈 Project Status

Alikel is under active development. Recent improvements include:

  • Seamless integration between Leif's task creation and the visual Task Management/Calendar page.
  • Enhanced task management system with calendar view and bulk actions.
  • Improved multi-agent conversation flow and reduced AI hallucination.
  • Implementation of detailed fitness profile tracking with measurement history.
  • Refined mobile UI and navigation.

We are continuously iterating based on user feedback and our long-term vision.

📄 License

Undecided ! it depends haha. Currently All Rights Reserved.

👨‍💻 Created By

Developed by Jordan.M / Alikel at Alikearn Studio.

See more projects at our portfolio: https://pixelle3-alikearn.com/portfolio

Visit alikel.net to try the platform!