DOCSEARCHbyAlgolia
Submit Docs

Awesome LLM Apps

repository·main·Indexed 13 days ago

https://github.com/shubhamsaboo/awesome-llm-apps

A collection of over 100 open-source AI agents, agent skills, and RAG (Retrieval-Augmented Generation) applications. It features end-to-end tested templates compatible with LLMs including Claude, Gemini, GPT, DeepSeek, Llama, and Qwen, with specific implementations for autonomous game playing agents, adaptive research teams, and competitor intelligence tools.

Tokens
208K
Snippets
595
Records
999
Agent score
99%

What's inside Awesome LLM Apps

  1. Overview of Generative UI and Agentic Frontends

    main

    Generative UI (Gen UI) allows LLMs to emit rich, interactive frontend components rather than just plain text. In these applications, the model determines the appropriate UI to display, and the frontend renders actual functional components that users can interact with (click, edit, respond). This closes the loop between the agent's reasoning and the user interface.

    This repository provides templates for several common Gen UI implementation patterns:

    • AG-UI / CopilotKit: Uses a streaming agent ↔ UI protocol designed for React applications.
    • Vercel AI SDK: Utilizes streamUI and React Server Components to generate UI.
    • LangChain / LangGraph UI: Renders structured tool calls as interactive components.
    • Custom tool-call → component renderers: Minimal DIY patterns for implementing Gen UI in any framework by mapping tool calls to specific component renders.

  2. Overview of AI Customer Support Agent Features

    main

    The AI Customer Support Agent provides the following capabilities:

    • Chat Interface: A Streamlit-based UI for interacting with the agent.
    • Persistent Memory: Uses Mem0 and Qdrant to remember customer profiles and past interaction history.
    • Synthetic Data Generation: Includes capabilities to generate synthetic data for testing and demonstration purposes.
    • Intelligent Reasoning: Powered by OpenAI's GPT-4o model.

  3. Overview of the AI Travel Agent with Memory

    main

    The AI Travel Agent is a Streamlit-based application that provides a chat interface for travel planning. It is designed to maintain context and remember user preferences across different interactions.

    Core Technologies

    • LLM: OpenAI's GPT-4o for intelligent response generation.
    • Memory Management: Mem0 for managing user preferences and conversation history.
    • Vector Database: Qdrant for persistent memory storage and retrieval.

    Key Capabilities

    • Persistent Memory: Remembers user preferences and past conversations using Mem0.
    • Chat Interface: A dedicated Streamlit UI for real-time interaction.
    • Memory Inspection: Allows users to view their specific conversation history and stored memory.

  4. Overview of the RAG Agentic System

    main

    This system is a Retrieval-Augmented Generation (RAG) agentic implementation that uses the following stack:

    • LLM: Cohere's Command-r7b-12-2024 model for chat and RAG tasks.
    • Embeddings: cohere embed-english-v3.0 model.
    • Vector Storage: Qdrant Cloud for storing document embeddings.
    • Orchestration: LangGraph (using create_react_agent) to manage agentic workflows.
    • RAG Framework: Langchain.
    • Web Search: DuckDuckGo (DuckDuckGoSearchRun) used as a fallback tool when document retrieval fails to provide relevant context.

    Key Capabilities

    • Document Processing: Upload PDFs, which are automatically chunked, embedded, and stored in Qdrant.
    • Intelligent Querying: Performs similarity searches with threshold filtering and provides source attribution.
    • Web Fallback: If no relevant documents are found in the vector store, the agent automatically switches to web research via DuckDuckGo.

  5. Overview of Resume & Job Matcher functionality

    main

    The Resume & Job Matcher is a Streamlit-based application designed to help job seekers optimize their resumes. By uploading both a Resume and a Job Description, the application uses an LLM to generate:

    • Fit Score: A percentage (0–100%) indicating how well the resume matches the job.
    • Strengths: Highlights of existing skills and experiences in the resume that align with the job.
    • Improvements: Specific, tailored suggestions to modify the resume to better fit the job description.

  6. Overview of the AI SEO Audit Team

    main

    The AI SEO Audit Team is an autonomous, multi-agent workflow built with Google ADK. It automates the process of auditing a webpage's SEO by crawling the URL, researching SERP (Search Engine Results Page) competition, and generating a prioritized optimization report in Markdown format.

    Core Workflow

    The system uses a SequentialAgent pattern where state is passed between three specialized agents:

    1. Page Auditor Agent: Uses firecrawl_scrape to inspect page structure and infer target keywords.
    2. Serp Analyst Agent: Analyzes Google Search results to identify competitor patterns and PAA (People Also Ask) questions.
    3. Optimization Advisor Agent: Synthesizes audit and SERP data into a final actionable report.

    Key Technologies

    • Google ADK: Orchestrates the multi-agent workflow.
    • Firecrawl via MCP (Model Context Protocol): Handles accurate page scraping.
    • Gemini 2.5 Flash: Powers the analysis and reporting logic.

  7. Overview of TripCraft AI

    main
    TripCraft AI is a multi-agent AI system designed to automate travel planning. It transforms simple user inputs (destination, dates, budget, travel style, and preferences) into complete, day-by-day travel itineraries. The system uses specialized agents working in parallel to handle different aspects of travel such as flights, hotels, activities, and budgeting.

  8. Overview of the AI Health & Fitness Planner Agent

    main

    The AI Health & Fitness Planner is a personalized agentic application built using the Agno AI Agent framework. It utilizes two specialized agents to provide comprehensive wellness guidance:

    • Health Agent: A specialist focused on providing dietary advice and generating detailed meal plans (including breakfast, lunch, dinner, snacks, hydration, and electrolytes) based on dietary preferences like Keto or Vegetarian.
    • Fitness Agent: A specialist focused on workout routines, covering warm-ups, main exercises, and cool-downs based on user fitness goals.

    The application accepts user inputs such as age, weight, height, activity level, dietary preferences, and fitness goals to generate tailored plans and supports interactive Q&A for follow-up questions.

  9. Overview of the AI Startup Trend Analysis Agent

    main

    The AI Startup Trend Analysis Agent is a tool designed for entrepreneurs to identify nascent trends, market gaps, and growth opportunities in specific sectors.

    Core Functionality

    • User Prompting: Users input specific startup sectors or technologies to research.
    • News Collection: Uses DuckDuckGo to gather recent startup news, funding rounds, and market analyses.
    • Summary Generation: Uses Newspaper4k to extract and generate concise summaries of verified information.
    • Trend Analysis: Leverages Claude 3.5 Sonnet to identify emerging patterns in funding, technology adoption, and market opportunities.
    • User Interface: Provides a web-based interface via Streamlit.

  10. Overview of AI Meme Generator Agent features

    main

    The AI Meme Generator Agent is a browser automation tool that uses AI agents to interact with imgflip.com. It automates the process of finding meme templates, inserting text, and extracting the resulting image links.

    Core Capabilities:

    • Multi-LLM Support: Supports Claude 3.5 Sonnet, GPT-4o, and Deepseek v3 with automatic model switching based on API key validation.
    • Browser Automation: Direct interaction with meme templates, automated template searching, and dynamic caption insertion.
    • Smart Workflow: Extracts action verbs from prompts, performs metaphorical template matching, and includes a multi-step quality validation and automatic retry mechanism.
    • Interface: Includes a model configuration sidebar, API key management, and direct meme previews with clickable links.

  11. Overview of Chat with PDF features

    main

    The Chat with PDF application is a lightweight RAG (Retrieval Augmented Generation) implementation that provides the following capabilities:

    • PDF Upload: Users can upload local PDF documents into the application.
    • Contextual Questioning: Users can ask natural language questions specifically about the uploaded PDF content.
    • RAG-driven Answers: The system retrieves relevant snippets from the PDF to provide accurate, context-aware answers using a selected LLM.

  12. Overview of the AI Research Agent with Memory

    main

    The AI Research Agent with Memory is a Streamlit-based application designed to assist users in searching for academic papers on arXiv. It goes beyond simple search by maintaining a persistent memory of user interests and past interactions to provide a more personalized research experience.

    Core Technologies:

    • LLM: OpenAI's GPT-4o-mini for processing and summarizing search results.
    • Web Browsing: MultiOn for navigating and retrieving web content.
    • Memory Management: Mem0 combined with Qdrant (vector database) for storing and retrieving user context and long-term interests.