Filters

AI Models

Stable DiffusionRedisQdrantChatGPTGPT-3CodexDALL-E-2AI21 LabsWhisperAlpacaLLaMAAuto-GPTGPT-5Cohere Coral

Data

RedisQdrant

Platforms

StreamlitReplitVercel

Discover Best Apps according to The Community!

Discover the most voted AI applications that have been built during lablab.ai hackathons. Explore our AI Hackathon projects and get inspired to create your own!

AllTrending 🚀Recent Winners 🥇Most votes 🔝
Sight beyond Sight

Sight beyond Sight

Our website enhances online content accessibility for the visually impaired with a cost-effective text-to-speech service using contemporary AI tools. Current market solutions lack necessary amenities and are costly. Working on the website: > once the website loads, the user inputs the URL of the website to be analyzed > this website is parsed using Beautiful Soup to gather the meaningful text content available on the page > this content is passed to the OpenAI text-davinci-003 model as a prompt and a summary is generated for the same > this summary is read out to the user using Azure in natural human tone > next, the website is again parsed using Beautiful Soup with the aim to download relevant images on that website > these images are then analyzed using Google Cloud Vision API and feature labels describing the prominent objects/contents of that image are generated > these labels are passed as a prompt to the OpenAI text-davinci-003 model and a meaningful sentence is generated which describes the images > the prompt already includes a set of sample labels and outputs that the model can use to understand the format of the desired output. > the image description generated in the above step is then read aloud using Azure. For Redis: Redis caches URL results for up to 3 hours, if URL exists in cache, output is displayed/read aloud. Otherwise, website is processed for new output. Results are removed after 3 hours for possible content changes. It allows for fast data access making it suitable for high performance use cases. For voice control: > using space bar, user can ask queries regarding summary through available chatbot > above query is converted to text via speech recognition library of python > this text and the summary are given to the OpenAI text-davinci-003 model as a prompt and the query is resolved > the result is spoken out and if speech unrecognized, an error message stating to retry is read aloud

Galacticos
RedisGPT-3
LectureMind

LectureMind

Once CoVid took over, everything was available on the internet. As industries and large-scale organizations fit their processes inside small laptop screens, the education sector had a tough time as well, especially trying to figure out the best possible ways to conduct online learning. Availability of physical resources, lack of technical training and expertise among faculty, and unpreparedness are just some of the reasons why online learning still is a huge challenge. We try to make things simpler by bringing in LectureMind. LectureMind takes care of a lot of things. All one has to do is upload an audio recording of a lecture into the platform and it: - Transcribes it and creates a summary - Generates a 10 point Quiz for every lecture - Generates 10 flash-cards for every lecture The idea is to simplify the process of online learning, make it fun for the students, have them play around with things a little and explore more and deep dive into things, because that's the best way to learn.

Neural Navigators
Vercel
application badge
LegacyLink AI

LegacyLink AI

LegacyLink AI is an AI-assisted legacy database modernization tool built for the IBM Bob Hackathon. It helps developers convert messy legacy SQL schemas into clean Python SQLAlchemy ORM project scaffolds, reducing the time needed to understand old database structures and generate boilerplate code. The system lets users upload a legacy SQL file, then automatically parses database objects such as tables, columns, data types, indexes, constraints, views, and functions. It normalizes cryptic names into readable conventions, such as converting tbl_CUST_MSTR_2012_v2 into Customer, fk_str_id into store_id, and dt_upd_dt into updated_at. LegacyLink AI generates a downloadable project containing SQLAlchemy 2.0 ORM models, database setup code, pytest test files, requirements.txt, README documentation, and a modernization report explaining the transformation. The generated project can be tested immediately using pytest. The project also includes an optional IBM watsonx.ai AI Assistant powered by IBM Granite. The assistant helps users explain schemas, summarize modernization reports, identify possible foreign-key-like fields, and understand generated ORM code in simple English. For enterprise traceability, AI Assistant interactions can be logged into PostgreSQL. IBM Bob can then connect through MCP to inspect the audit logs and summarize recent assistant activity, including the user question, model used, SQL file processed, table count, and status. IBM Bob was used as the development partner for planning, coding, debugging, testing, documentation, and MCP workflow validation. The final result combines Streamlit, SQLAlchemy, pytest, IBM watsonx.ai, PostgreSQL, MCP, and IBM Bob into a practical modernization workflow for legacy database teams.

Try-Tri-Hard
Streamlit
application badge
IBM watsonx AssistantIBM GraniteIBM+2
RepoMind : AI Onboarding Copilot

RepoMind : AI Onboarding Copilot

RepoMind solves one of the biggest pain points in software development: onboarding into large and complex codebases. Developers joining new projects often spend weeks trying to understand system architecture, dependencies, and workflows. This slows productivity, delays delivery, and creates frustration across engineering teams. RepoMind is an AI-powered onboarding copilot built using IBM watsonx and IBM Granite. Users paste a GitHub repository URL and instantly receive an architecture summary, module breakdowns, execution flow explanations, and a conversational AI assistant that answers questions like “How does authentication work?” or “Where is the payment logic handled?” The platform uses IBM Granite 3 8B Instruct through watsonx.ai to deliver contextual understanding of entire repositories. The backend is powered by FastAPI and GitPython for repository analysis, while the frontend runs on Streamlit for a simple and interactive experience. RepoMind was developed in collaboration with IBM Bob and is designed for junior developers, onboarding engineers, open-source contributors, and teams managing legacy systems. The business opportunity is significant. The global developer tools market exceeds $26 billion, while AI-powered coding and onboarding tools continue to grow rapidly. RepoMind follows a scalable SaaS model with free access for public repositories and premium plans for private repositories, team dashboards, and enterprise integrations with GitHub, GitLab, and Bitbucket. Unlike tools focused on autocomplete or code search, RepoMind specializes in full codebase comprehension through natural-language interaction, helping developers understand systems faster and reduce onboarding time from weeks to minutes.

Kernel Panic
Streamlit
application badge
IBM watsonx AssistantStreamlitIBM Granite+1
EasyFastAI - Empower your business with AI

EasyFastAI - Empower your business with AI

We Empower WhatsApp with AI in just 3 steps: 1 - Scan QR Code Utilize the WhatsApp-linked device option to scan the QR code generated by our system, facilitating the seamless connection of your WhatsApp number 2 - Train AI Begin the process of uploading documents to facilitate a deeper understanding of your business by our AI 3 - AI Ready For Use Our AI is now poised to efficiently manage customer interactions around the clock, leveraging the data provided in any language EasyFastAI can retrieve information from several types of documents, can reply to voice notes in several languages and can perform actions

Easy Fast AI
Assistants APIOpenAIGPT-3.5+6
CodeAtlas

CodeAtlas

CodeAtlas is an AI-powered engineering intelligence platform designed to help developers and engineering teams understand, audit, and modernize complex codebases. Modern software systems are increasingly difficult to reason about. As repositories grow, teams lose visibility into architecture, dependencies, technical debt, and change impact. Refactoring becomes risky, onboarding slows down, and modernization projects become expensive and unpredictable. CodeAtlas solves this problem by transforming GitHub repositories into interactive engineering intelligence. Users can import a repository directly from GitHub, and CodeAtlas automatically scans the codebase, analyzes dependencies, maps architecture relationships, identifies high-risk modules, and generates modernization insights. The platform includes: Interactive dependency graph visualization Risk-based module analysis Blast-radius simulation Technical debt estimation Architecture health scoring Dependency confidence analysis Engineering findings and onboarding guidance Suggested modernization sequences Exportable engineering reports One of the key goals of CodeAtlas is transparency. Instead of pretending the analysis is perfect, the system exposes dependency confidence limitations caused by unresolved imports, framework aliases, or dynamic imports. This makes the platform more trustworthy and operationally useful for real engineering teams. The frontend was built with Next.js, TypeScript, and React Flow, while the backend uses FastAPI with a custom repository scanner, dependency graph engine, and risk analysis system. CodeAtlas was built during the IBM Bob Hackathon to demonstrate how AI can assist engineering organizations with architecture visibility, modernization planning, and repository intelligence.

CodeAtlas
Vercel
application badge
ChatGPTIBMIBM watsonx Assistant
PRISM (Pull Request Intelligent Semantic Monitor)

PRISM (Pull Request Intelligent Semantic Monitor)

PRISM (Pull Request Intelligent Semantic Monitor) is a semantic risk-analysis layer that sits on top of the existing pull request workflow, catching the silent failures that traditional CI/CD pipelines cannot see. While conventional checks confirm that code compiles and tests pass, they remain blind to behavioral inconsistencies — renamed API fields, altered enum values, or refactored permission logic — that quietly break downstream services in production. PRISM closes this gap by combining static AST analysis, dependency graph traversal, and IBM Bob-powered reasoning to trace how each change propagates through the codebase. It delivers a risk score, impacted-component map, and targeted regression scenarios through a CLI summary and interactive dashboard, answering the question CI cannot: should this code merge?

Tech Cats
IBM watsonx AssistantIBM Granite
Trellis: The Knowledge Fabric for Law Firms

Trellis: The Knowledge Fabric for Law Firms

Every law firm sits on an unmined fortune: the experiential wisdom of senior partners. Yet, when a veteran attorney retires or departs, decades of strategic instinct permanently evaporate. This institutional knowledge loss costs mid-to-large firms $15M to $40M annually. The barrier to documentation is economic. Lawyers bill in six-minute increments; spending time writing down lessons represents an immediate loss of billable revenue. While generative AI should solve this, strict obligations toward attorney-client privilege block public LLM tools. Meanwhile, vertical legal tools like Harvey provide generic legal advice but lack a firm's specific internal wisdom. Trellis solves this with a secure, two-tier system architecture. The first layer is the Personal Second Brain. This is a private, on-device edge environment running Gemini Nano where attorneys quickly capture unstructured thoughts via voice memos, notes, or image OCR. Because data stays local, capture is entirely unredacted and factual. The second layer is the Team-Managed Knowledge Graph. This shared ledger is built through an automated, dual-pass sanitization pipeline. First, Microsoft Presidio strips explicit PII like names, dates, and entities. Second, Gemini Pro abstracts the specific case details into generalized strategic principles. The lawyer reviews a side-by-side diff before approving publication to the firm's shared memory. Using a hybrid vector-graph RAG model, team members can query this collective intelligence using natural language. To eliminate legal risk, Trellis enforces a hard deterministic guardrail: if the retrieved context score falls below a strict threshold, the system executes a strict refusal rather than risk a hallucination. Finally, Trellis exposes a Model Context Protocol (MCP) endpoint. This allows third-party tools like Harvey, CoCounsel, or Copilot to securely plug directly into the substrate, grounding generic legal AI capabilities in the firm's actual historical wisdom.

Pan sa Manila
medal
Vercel
application badge
Gemini AIGemini 3 proGemini 3 Flash+2
ARCA SENTRY

ARCA SENTRY

Thousands of companies are racing to ship AI — chatbots, voice agents, automated customer service. But who audits them? The EU AI Act takes full effect in August 2026, with fines reaching up to 7% of global revenue. A single non-compliant chatbot response can trigger an investigation. Most teams have no defense. ARCA SENTRY is the compliance brain for enterprise AI. A council of five specialised auditor agents — running on DeepSeek through Featherless — audits every interaction your AI produces, in real time, against the full EU regulatory stack: EU AI Act, GDPR, DORA, PII leakage and prompt injection. It integrates with any AI channel: drop-in proxy mode (OpenAI / Anthropic / Gemini), raw HTTP endpoints, WhatsApp Business, Facebook Messenger. Zero code changes beyond the base URL. Before you ship, run the Red Team: an automated pen-test suite that probes your agents against every European framework and exports forensic reports in PDF, Markdown and HTML — showing exactly how the model can be broken. In production, the gateway intercepts violations mid-flight, blocking the response before it reaches the user. Every incident auto-generates a remediation ticket with estimated fine, suggested fix (generated by Gemini Pro), and a tamper-evident SHA-256 hash chain for forensic reproducibility. Auto-detects five languages: Italian, Portuguese, Spanish, Chinese, English. Don't let AI become your most expensive line item. Take back control of your agents. The council of auditors guards your AI — and your business. ARCA SENTRY. Don't pay the fine. Pay attention.

AI Society - B Drive
Vercel
application badge
Claude CodeGemini AIFeatherless+6
MyThorch

MyThorch

MyThorch is a revolutionary document interaction app that leverages user behavior to embed vectors and store them in a Redis vector database. This process creates a long-term memory AI that truly understands the user's needs and preferences, leading to a highly efficient reading experience. The user uploads documents and the app filters for key points and also provide content the user cares about. The user can then select any part of the document to clarify further and track down references for specific parts of the generated document, ensuring accuracy and avoiding AI hallucinations. Using React for the frontend and Flask for the backend, we deliver a fast and efficient user experience. The GPT-3 API generates personalized documents based on user focus, while Redis stores previous interactions to reduce token input to GPT-3. Constant data collection helps us adapt to the user's needs for a personalized experience. Future plans include integration with Kindle, Kobo, Google, and Firefox, as well as partnerships with document management systems and cloud storage platforms. Our subscription-based business model offers premium features for $2/month. MyThorch is a game-changer for students, teachers, educators, and researchers, offering enhanced productivity and time-saving benefits.

HCMUwUS
medal
Vercel
application badge
RedisCohere GenerateGPT-3
BiteBuddy Meal Planner

BiteBuddy Meal Planner

BiteBuddy is a meal-planning tool that is prized for its customization. What this entails is that for every different user, a different meal plan is generated based on the user's preferences, needs, and dislikes. A tool to create a meal plan needs to understand what the customer needs. Therefore, the tool asks the user about their body characteristics to help organize the facts needed to create a balanced diet. Precisely, the user's Body Mass Index (BMI) is calculated to give an idea about the possible goals and hence, display an adequate diet. After that, the tool asks the user about any allergies, if any exist, so as not to include these foods in the plan. Moreover, they are asked to input any preferences they might have to include in their diet. The user input is then processed to generate the data. After which, the data is sent via the APIs through properly formatted prompts. The generated meal plan is then displayed in table format for every day of the week and at least three meals per day. Each meal has its nutritional values displayed, including carbohydrates, calories, fats, and proteins; the tool also suggests the right portions and the recipe with the ingredients that aid in preparing that meal. The recipe sub-window also illustrates the meal by providing a photo. Sorting is a feature that can be used to sort the meals based on their nutritional facts. The main advantage of this website is that it groups everything that the user might need to plan a healthy diet, considering their allergies and any food preferences they might have for the week. The tool replaces the need to visit nutritionists and explore different food regimes. It also replaces searching and exploring various recipes among multiple cooking platforms.

The DeCoders
ChatGPTDALL-E-2GPT-3
InsightStream: AI Customer Analytics

InsightStream: AI Customer Analytics

InsightStream is an AI-powered customer business analytics dashboard designed to help businesses transform raw customer data into meaningful insights and smarter decisions. Many organizations collect large amounts of customer data but struggle to analyze it effectively due to the complexity of traditional analytics tools. InsightStream solves this problem by providing an interactive and user-friendly platform that makes customer analytics accessible to both technical and non-technical users. The platform enables users to upload customer datasets, explore interactive dashboards, perform RFM (Recency, Frequency, Monetary) customer segmentation, identify churn risks, and generate actionable business recommendations to improve customer retention and customer lifetime value. It also supports SQL-based data exploration, CSV uploads, SQLite database integration, and exportable visual reports for better decision-making. Built using Python, Streamlit, SQLite, Pandas, and Plotly, InsightStream focuses on scalability, modular development, and rapid prototyping. Throughout the development process, IBM Bob served as an AI-powered development partner by assisting with code generation, debugging, architecture planning, workflow optimization, and repository understanding. InsightStream demonstrates how AI-assisted development can accelerate the creation of practical, business-focused software solutions with real-world value.

Token Overflow
Streamlit
application badge
IBM
Autonomous Agents from APIs - Zero Code Builder

Autonomous Agents from APIs - Zero Code Builder

Using Vultr, Groq, Llama, Coral, and Fetch.ai, our project introduces a transformative way to generate and deploy Autonomous AI Agents from APIs—a no-code, prompt-driven AI Agent Builder and conversational interface to manage intelligent agents across enterprise systems in real time. From Prompt ➜ Plan ➜ Code ➜ Execute ➜ Iterate ➜ Deploy ➜ Use (on Coral, Custom, and Fetch AI) User explains their goal in natural language. Our AI locates relevant APIs, generates step-by-step agent plans, writes the code, deploys tools to Vultr-hosted MCP servers, and registers them with Coral and Fetch.ai. All agents are accessed via a multi-modal chat interface—text or voice—enabling users to test, refine, and operate agents live. These agents interact with real-world APIs via the MCP servers created through the Agent Builder. No code. Just intent. From developers and marketers to support and operations teams—anyone can turn apps into automations, tools, or digital workers. Enterprise-ready and built for the future of work, Vortex IQ enables users to go from idea to working agent in minutes. All AI inference runs on Groq’s ultra-low-latency hardware using Llama 3 models, making it fast, private, and cost-efficient. Agents are also composable via Fetch.ai’s Agentverse and ASI:One. Future-Ready with Scalable Impact: a. 3,000+ Retailers: Rolling out AI agents across BigCommerce, Adobe Commerce, Shopify, and Salesforce Commerce Cloud. b. Composable Agent Hub: 100s of reusable agents and templates for plug-and-play use across any API-powered system. c. Cloud Partnerships: Vultr, GCP, and Azure hosting for global agentic compute-as-a-service access. Tech Stack: Hosted on Vultr, using Groq, Llama 3, Coral Protocol, and Fetch.ai for compute, orchestration, and autonomy. Other tech: Google ADK, Next.js From prompt to production—this is how the future of work is built.

Autonomous Agents from APIs - Vultr Track
medal
GroqLlama 3.2Llama 3.1+4
CIPHER AI

CIPHER AI

Cipher is an intelligent workflow orchestration platform designed to simplify how people build, automate, and optimize complex processes. Instead of relying on rigid automation rules, Cipher combines the power of AI-driven agents, real-time collaboration, and adaptive decision-making to create workflows that can think, respond, and evolve. Users can visually design workflows, connect specialized AI agents, and monitor every step of execution through a unified interface. Cipher not only automates repetitive tasks but also detects errors, provides contextual assistance, and recommends improvements to enhance efficiency and reliability. Whether it's streamlining business operations, accelerating software development, managing content pipelines, or coordinating data-driven tasks, Cipher adapts to diverse use cases without requiring extensive technical expertise. Built with a focus on security, scalability, and usability, Cipher empowers individuals and teams to transform ideas into intelligent processes that deliver meaningful outcomes. By bridging the gap between traditional automation and autonomous AI collaboration, Cipher enables organizations to work faster, make better decisions, and unlock new levels of productivity in an increasingly complex digital world.

Runtime Error
Vercel
application badge
Anthropic ClaudeAntigravityChatGPT+9
Novus - Research Intelligence Assistant

Novus - Research Intelligence Assistant

Novus is a multi-agent AI platform that transforms enterprise R&D workflows by automating the entire research intelligence pipeline — from duplication detection to grant proposal generation — in under 3 minutes. The Problem: Fortune 500 companies lose $28 billion annually funding research that already exists. R&D teams spend months writing grant proposals across disconnected tools with no unified system to check duplication, find grants, and ensure compliance simultaneously. Our Solution: Novus deploys 6 specialized AI agents coordinated through Band SDK: 1. Intake Agent — Parses R&D proposals and extracts competitive intelligence signals 2. Duplication Scout — Scans 200M+ papers across Semantic Scholar, OpenAlex and arXiv in real time 3. Relevance Agent — Scores proposals against current industry funding trends 4. Eligibility Agent — Matches to live grants from Grants.gov and NIH Reporter, surfaces already-funded competing research 5. Proposal Writer — Drafts complete, funder-aligned grant proposals automatically 6. Compliance Agent — Reviews against grant requirements before submission Key Features: - Real academic paper search with authors, years and source links - Live grant matching with direct Apply Now links - Already-funded similar research discovery with award amounts - International funder recommendations including EU Horizon, Wellcome Trust and Gates Foundation - Full audit trail via AgentOps - Beautiful React Flow live agent pipeline visualization Tech Stack: Band SDK, Anthropic Claude, AI/ML API, Featherless AI, Semantic Scholar, OpenAlex, arXiv, Grants.gov, NIH Reporter, Next.js, React Flow, AgentOps, Qdrant

Novus
AgentOpsAI/ML APIAnthropic Claude+3
AIYu

AIYu

Supercharge your business operation by using AI technology. We support small business by introducing smart automation into their daily business operation. We leverage different open AI stack and Redis in our implementation. We achieved 10x faster operation and manage to demo our product to our potential first customer.

AI.Yu: Supercharge your business operation
Reinforcement LearningRedisWhisper+1
Vela — AI Agency Command Center

Vela — AI Agency Command Center

AI agencies deploy agents for clients — but once they go live, there's zero visibility. No monitoring. No audit trails. No way to know if agents are working or failing. Vela is the command center built for AI agencies. WHAT VELA DOES: → Monitor every client and deployed agent in real time → Receive live events via webhooks from any platform (n8n, Vapi, Voiceflow, Make) → Track connection status with source IP detection and audit trails → Generate AI-powered client performance reports in one click → Enterprise security powered by Veea Lobster Trap — every AI interaction inspected for prompt injection, PII exfiltration, and risk scoring SECURITY LAYER (Track 1 — Veea): Every report generation goes through Lobster Trap proxy. Real-time risk scoring, intent mismatch detection, full audit logs visible at tryvela.io/security AI LAYER (Track 2 — Google): Reports powered by Google Gemini 2.5 Flash via OpenAI-compatible endpoint TECH STACK: Python + Flask, Google Gemini 2.5 Flash, Veea Lobster Trap, Vultr Cloud, Webhook API, SSL/HTTPS Live at: https://tryvela.io Demo: [email protected] / demo1234

Vela
Gemini AIGemini 3 FlashGenerative AI Studio+4
CogniCore

CogniCore

CogniCore is an AI-powered enterprise platform that transforms meetings into production-ready operational workflows. Instead of relying on manual documentation, fragmented communication, and multiple disconnected tools, CogniCore automatically captures conversations, understands business and technical context, extracts requirements, identifies risks, and generates actionable tasks in real time. Through its integrated systems, CogniScribe and CogniOPS, the platform orchestrates planning, governance, development workflows, and AI-assisted execution from a single interface. By reducing miscommunication, accelerating project delivery, and minimizing operational overhead, CogniCore enables organizations to move seamlessly from discussion to implementation with speed, accuracy, scalability, and enterprise-grade automation.

ChamPeng
AMD ROCmAnthropic ClaudeAMD Developer Cloud+14
ICDPath - Your Path to Understanding Your Health

ICDPath - Your Path to Understanding Your Health

ICDPath is an innovative AI-powered application designed to bridge the gap between complex medical information and patient understanding. Our solution offers two primary functionalities: For users with medical reports, ICDPath analyzes the document, identifying and explaining ICD codes, medical terminology and medications. Users can then engage in a QnA session to gain deeper insights into their health conditions, treatment options and potential side effects. For users without medical reports, ICDPath provides a symptom-based analysis, suggesting possible ICD codes and conditions. This feature helps users understand potential diagnoses and offers general health advice. Built using Langflow, ICDPath aims to empower users with personalized health information. Our app translates medical jargon into easily understandable language, enabling users to take control of their health journey. We're currently developing a mobile version to make this valuable tool even more accessible. ICDPath is not just a simple app, it's a personal health assistant, designed to enhance health literacy and patient empowerment.

Astres
Streamlit
application badge
StreamlitlangflowGPT-4 Vision+1
multiLLM Collaboration

multiLLM Collaboration

The Multi-LLM Collaboration System revolutionizes AI interaction by creating expert teams where different AI models assume specialized professional roles for collaborative discussions. Unlike traditional single-model conversations, this platform harnesses the unique strengths of GPT-5, Claude, Gemini, and other models through structured, multi-round discussions. The system features intelligent coordination using AI-powered workflow management, automatic role clarification, and consensus detection. Each AI participant takes on specific expertise (CEO, researcher, analyst, etc.) to provide diverse perspectives on complex topics. The platform supports bilingual operation (English/Chinese) and generates comprehensive reports with actionable insights. Applications span across medical research discussions, academic collaborations, business strategy sessions, and government policy meetings. The round-based discussion format ensures thorough analysis while the intelligent coordinator determines optimal progression and stopping points. Users can export complete transcripts and structured documents, making it ideal for professional environments requiring documented collaborative analysis. This innovative approach transforms AI from individual assistants into collaborative expert teams, delivering insights greater than the sum of their parts.

ThePowerPower
OpenAIGemini AIAnthropic Claude+3
GRUF

GRUF

Reduced Food Costs and Dependency: By empowering individuals to grow their own food, governments can indirectly reduce the strain on food supply chains, potentially leading to lower food prices and reduced reliance on external sources. Improved Public Health: Homegrown, organic produce can contribute to healthier communities, reducing healthcare costs associated with diet-related illnesses. Community Resilience: Urban gardening can foster a sense of community, strengthen social bonds, and create more resilient neighborhoods, which can indirectly reduce the burden on government services. New Social Programs may be created to promote the planting, harvesting and communal exchange of urban food, providing access to seeds and other resources. Thus, our GRUF App shall play a very important role, leveraging the power of xAI Grok to help us advance towards a much healthier and stronger economy.

GRUF
Vercel
application badge
GrokVercel
Code PRO

Code PRO

The code optimization and error fixing app is a powerful tool for developers and programmers that is designed to help them identify and fix errors in their code, as well as optimize it for better performance. Built using Codex, a cutting-edge AI language model, the app supports around 200 different programming languages, making it a versatile and comprehensive solution for all kinds of developers. The app provides a user-friendly interface that allows developers to easily upload their code and quickly identify any errors or issues. The app uses advanced algorithms and machine learning techniques to analyze the code and highlight any errors or bugs that may be present. It also provides suggestions and recommendations for fixing the errors, as well as optimizing the code for better performance. One of the key features of the app is its ability to provide real-time feedback as the developer is coding. This means that as the developer writes the code, the app can identify errors and provide suggestions for how to fix them. This can save developers a significant amount of time and effort, as they no longer have to spend hours debugging their code manually. The app also provides a range of other useful features, including code formatting, syntax highlighting, and autocomplete functionality. These features make it easier for developers to write code that is easy to read and understand, and that follows best practices and coding standards. Overall, the code optimization and error fixing app is a powerful and comprehensive tool that can help developers to write better code, faster. Whether you are a beginner or an experienced developer, this app can help you to identify and fix errors in your code, and optimize it for better performance.

BitsPeeps
Streamlit
application badge
Codex
Vela — AI Agency Command Center

Vela — AI Agency Command Center

AI agencies deploy agents for clients — but once they go live, there's zero visibility. No monitoring. No audit trails. No way to know if agents are working or failing. Vela is the command center built for AI agencies. WHAT VELA DOES: → Monitor every client and deployed agent in real time → Receive live events via webhooks from any platform (n8n, Vapi, Voiceflow, Make) → Track connection status with source IP detection and audit trails → Generate AI-powered client reports in one click — powered by Google Gemini 2.5 Flash → Enterprise security via Veea Lobster Trap — prompt injection detection, risk scoring, full audit trails TRACKS: → Enterprise Utility — solves a real problem every AI agency faces → Agentic Workflows — webhook-based real-time agent event system → Vultr Award — fully deployed on Vultr Ubuntu 24.04 TECH STACK: Python + Flask, Google Gemini 2.5 Flash, Veea Lobster Trap, Vultr Cloud, Webhook API, SSL/HTTPS Live at: https://tryvela.io Demo: [email protected] / demo1234 GitHub: https://github.com/proresin382-cpu/vela

Vela
Gemini AIGemini 3 FlashVULTR+2
HireMate

HireMate

HireMate is an AI-powered multi-agent recruitment platform that transforms the traditional hiring pipeline into a fully autonomous, intelligent workflow. Instead of relying on manual resume screening and subjective evaluations, HireMate deploys five specialized AI agents — Orchestrator, Resume Analyzer, Screening Evaluator, Task Grader, and Final Judge — that collaborate in real-time through Band multi-agent orchestration, powered by NVIDIA Nemotron 550B large language model. When a candidate submits an application, the Orchestrator creates a dedicated evaluation room and triggers the pipeline. The Resume Analyzer parses the resume, extracts skills, and scores job fit. The Screening Evaluator generates role-specific questions and evaluates responses. The Task Grader assesses technical submissions against structured rubrics. Finally, the Final Judge aggregates all agent scores, applies weighted criteria, and delivers a definitive SELECT or REJECT decision with a detailed breakdown — all within 60 seconds. Built with Next.js, Express.js, Prisma ORM, and PostgreSQL, the platform features separate admin and student dashboards, real-time agent pipeline visualization, and automated interview scheduling. Deployed as three independent microservices on Railway, HireMate reduces manual screening time by 85% and eliminates unconscious hiring bias through consistent AI evaluation.

Agent Hunterzzz
AntigravityClaude CodeGPT-5+1
AEGIS Fund: Governance No Trader Can Override

AEGIS Fund: Governance No Trader Can Override

very catastrophic hedge fund failure — LTCM, Archegos, FTX — shares one root cause: the risk function was structurally powerless, overruled by the people making money. AEGIS Fund inverts this. It's a fully autonomous hedge fund run by 11 AI agents — a CIO, competing portfolio managers, traders, sealed strategy bots, and independent CRO, Compliance, and Research officers — each on a different framework (LangGraph, NVIDIA NIM / Llama 3.3 70B, and deterministic adapters), coordinating entirely through Band. Band is the collaboration layer, not a wrapper. Every directive, report, advisory, block, and halt is a Band message, routed by @mention through shared rooms and permanently logged as a unified audit trail. No agent calls another directly; remove Band and the system cannot function. The command chain flows top-down (capital mandates) and bottom-up (aggregated reports), while the CRO and Compliance agents cut across it with hard-coded authority no portfolio manager or CIO can bypass. The strategies stay sealed — the governance layer supervises bots it cannot see, with only abstract risk signals crossing the boundary. In the live dashboard, the fund runs eight market ticks: at tick 5 Compliance blocks a desk on a rules violation; at tick 7 the CRO halts the entire fund on a drawdown breach — two independent overseers, different mandates, with zero human intervention. AEGIS is more than a finance demo. It's a blueprint for any high-stakes workflow — lending, claims, compliance review — where oversight exists on paper but bends under pressure. It shows that multi-agent systems can have enforceable, auditable authority structures, coordinated across frameworks through Band.

Agentic Avengers
AntigravityBand Control PlaneBand Intergrations+2
Sense AI

Sense AI

Sense AI is an innovative assistive technology project aimed at breaking communication barriers and empowering individuals with disabilities. The solution combines multiple accessibility tools into a single intelligent platform. For the deaf, it offers real-time voice-to-text transcription, enabling them to follow conversations seamlessly. For the mute, it provides text-to-speech conversion, allowing them to communicate effectively in daily interactions. Beyond communication, Sense AI leverages artificial intelligence to deliver personalized support and plans to integrate IoT-based devices to assist with mobility and independence. By uniting these technologies, Sense AI creates an inclusive ecosystem that promotes accessibility, equal opportunities, and improved quality of life.

Sense AI
Vercel
application badge
GPT-3.5
ScopeShield: Clarity Before Code

ScopeShield: Clarity Before Code

ScopeShield is a developer-focused tool designed to prevent scope creep before work begins. It transforms vague client or stakeholder requests into a clear technical scope contract, helping teams understand what is really being asked, what parts of the codebase may be affected, what risks are hidden behind the request, and what questions need to be answered before making commitments. Instead of reacting after a project grows out of control, ScopeShield helps developers, freelancers, agencies, and product teams clarify expectations early. The platform generates a structured analysis that includes hidden scope items, impacted areas, technical risks, clarifying questions, implementation steps, realistic time estimates, and a professional client-ready response. By turning ambiguity into actionable technical clarity, ScopeShield protects time, improves communication, and reduces costly misunderstandings.

Script Hunters
Vercel
application badge
IBMAI/ML APIAntigravity+2
ClinSight: Agentic Multimodal Clinical AI

ClinSight: Agentic Multimodal Clinical AI

Every year, 795,00 Americans are harmed by delayed diagnosis in emergency departments. Preliminary chest X-ray review takes 30-60 minutes. Rural hospitals wait 4-24 hours for teleradiology. ClinSight is an open-source multimodal clinical intelligence system built entirely on AMD hardware. It ingests chest X-ray images, lab values, vitals, and triage notes simultaneously — then reasons across all modalities through a compiled LangGraph agent pipeline on AMD Instinct MI300X via ROCm 7.0 and vLLM. Architecture: 5 parent agents orchestrate 7 subagents (12 reasoning nodes). Coordinator validates input and runs pediatric safety gates. Radiologist analyzes X-rays via Qwen2.5-VL-7B. Lab Analyst detects critical values and correlates patterns. Safety runs 3 parallel checks (contradiction, hallucination guard, bias audit) with a merge node. Documenter produces deterministic ESI scoring, differential diagnosis, and structured reports. Dual-Model Stack: Qwen2.5-VL-7B-Instruct (vision, ~14GB) + Qwen3.5-35B-A3B MoE (reasoning, ~70GB) = ~99GB / 192GB HBM3. Both models served simultaneously via vLLM on ROCm 7.0 — impossible on H100 80GB without quantization. Live Evidence: 50-case pure CXR benchmark on real MI300X. Mean latency: 23.02s. All 50 cases live, zero cache. GPU utilization: 10-49%, power 231-263W. rocm-smi evidence captured at baseline, during, and post-inference. Safety & Rigor: Physician-in-the-loop by design. Pediatric gate blocks adult-trained recommendations for under-18 patients. Bias auditor stratifies by age and sex. ESI scoring is rules-based, never LLM-generated. Apache 2.0 license.

Recourse
AMD Developer CloudAMD ROCmClaude Code+7
EquiPulse AI

EquiPulse AI

Millions of micro-businesses in emerging markets run their local economies but face substantial revenue losses due to an inability to predict demand, handle mathematical reconciliation, or access real-time retail intelligence. Existing ERP and POS networks are expensive, internet-dependent, and built for massive corporations rather than the small merchants of developing nations like Bangladesh. EquiPulse AI bridges this gap with a fully localized, offline-first business intelligence system engineered specifically for local merchants. Built with a simple, swipe-based user interface, the system stays operational even when internet access drops, shifting to local on-device fallback AI model pipelines without stalling business transactions. Technical Core and AI Architecture: 1. Intelligent Data Analytics: A user uploads a standard transaction CSV into the system. Our browser-based client-side DuckDB engine analyzes transactional data instantly, dynamically updating UI widgets and analytics components via a global POS context layer. 2. Cross-Framework Action Cards: Powered by Gemini 1.5 Flash inference, the system builds structured action cards for stock restocking, optimal price adjustments, and demand forecasting. These trigger instant automated workflows via webhooks. 3. Multimodal Receipt Parsing: A localized multimodal OCR engine scans physical Bangla invoices, automatically maps inventory relationships, and handles mathematical reconciliation. 4. Hybrid Graph RAG Layer: Instead of feeding huge unstructured logs into the models, our symmetric scrolling engine queries a custom hybrid Graph RAG pipeline. This captures intricate localized behaviors like massive seasonal demand shifts (e.g., Eid or Pohela Boishakh) for high-accuracy local inferences. EquiPulse AI operates on a tech stack combining Gemini AI, local Llama models, DuckDB, IndexedDB, and Firebase security rules. Furthermore, moving from a smart concept to a production-ready engine.

EquiSaaS BD
FeatherlessGemini AIAI/ML API
SentimentSense AI

SentimentSense AI

SentimentSense is an innovative AI-powered sentiment analysis app purposefully built for Monday.com's brand-new AI assistant. It offers accurate detection and analysis of emotions, specifically tailored to enhance the functionality of the Monday.com platform. By seamlessly integrating with Monday.com, SentimentSense allows users to leverage the power of AI to gain deep insights into sentiments expressed within their Monday.com items, groups, boards, and projects. In addition to its advanced sentiment analysis capabilities, SentimentSense goes beyond by providing assessment reports for selected users. These reports evaluate strengths, areas to improve, and a summary of individual users' performance within the Monday.com ecosystem. With SentimentSense, Monday.com users can effortlessly monitor and analyze customer feedback and public sentiment directly within the platform. This empowers businesses to make data-driven decisions, identify areas for improvement, and promptly respond to customer concerns, all within the familiar Monday.com environment. Utilizing advanced machine learning algorithms and natural language processing techniques, SentimentSense delivers sentiment analysis, enabling users to gauge public perception, measure brand sentiment, and uncover emerging trends. Its intuitive interface and interactive visualizations make interpreting and sharing sentiment insights seamless within the Monday.com ecosystem. In summary, SentimentSense is a purpose-built AI-powered sentiment analysis app designed to seamlessly integrate with Monday.com, providing Monday.com users with enhanced sentiment analysis capabilities to make informed decisions and gain valuable insights and suggested action items within their projects and workflows. The assessment reports further evaluate strengths, areas to improve, and an overall summary, enabling users to enhance their performance and drive better outcomes.

Appfire Team
medal
OpenAI
30apps loaded