Meta

Top Builders

Explore the top contributors showcasing the highest number of app submissions within our community.

General
Company	Meta Platforms, Inc.
Founded	January 4, 2004
Headquarters	Menlo Park, California, U.S.
Repository	https://github.com/facebook

Key Products and Research

Meta has developed a range of AI products designed to enhance various aspects of technology and user experience. Here’s a brief overview of these AI products:

LLaMA (Large Language Model Meta AI)

LLaMA is a series of large language models designed for natural language processing tasks. These models, including the latest LLaMA 3.1, are known for their advanced capabilities in text generation, understanding, and multilingual processing. They are available as open-source models, promoting innovation and research in AI Meta | Social Metaverse Company,Facebook.

Meta AI

Meta AI is an intelligent assistant integrated across Meta’s platforms, such as Facebook, Instagram, WhatsApp, and Messenger. Powered by LLaMA models, it helps users with tasks like content creation, information retrieval, and personalized interactions Meta | Social Metaverse Company.

PyTorch

PyTorch is an open-source machine learning library developed by Meta and widely used in both research and industry. It provides tools for building and training deep learning models and has become a standard framework in the AI community Facebook.

Meta AI Research (FAIR)

Meta’s AI research division, formerly known as FAIR (Facebook AI Research), focuses on advancing the field of AI through open research and collaboration. This division works on various AI challenges, including computer vision, natural language processing, and generative AI Facebook.

Meta AI in the Metaverse

Meta is also incorporating AI into its metaverse initiatives, using AI to create immersive experiences in virtual and augmented reality. This includes developing AI-driven avatars, enhancing virtual environments, and improving interaction within the metaverse Meta | Social Metaverse Company.

AI for Ads

Meta leverages AI to optimize ad targeting, delivery, and measurement across its platforms. AI algorithms analyze vast amounts of data to improve the effectiveness of advertising campaigns, making them more relevant to users and efficient for advertisers Meta | Social Metaverse Company.

LLaMA Impact Grants

The LLaMA Impact Grants program, launched by Meta, aims to support and encourage the innovative use of its LLaMA (Large Language Model Meta AI) models to address critical challenges in various sectors, including education, environmental sustainability, and public good. This initiative offers financial grants and resources to researchers, nonprofits, and other organizations that seek to leverage LLaMA models for impactful projects. The program highlights Meta’s commitment to responsible AI development and its belief in the potential of AI to drive positive social change.

For more details, visit the LLaMA Impact Grants page.

Edit on GitHub

Meta AI Technologies Hackathon projects

Discover innovative solutions crafted with Meta AI Technologies, developed by our community members during our engaging hackathons.

NEXUS AI trading agent

NEXUS is a trust-aware autonomous AI trading agent built for the next generation of financial agents that must do more than just trade. It was designed around a core challenge in AI finance: enabling agents to interact with capital safely, execute strategies autonomously, and demonstrate transparent, verifiable behavior. The project combines three strategy modes (algo, llm, hybrid), four risk profiles, DEX-based execution through a RiskRouter-compatible flow, ERC-8004 agent identity, and EIP-712 checkpoint logging. This means NEXUS is not just an automated trader, but a system built to make every important decision inspectable and accountable. One of its biggest strengths is trust infrastructure. NEXUS aligns with the ERC-8004 vision of identity, reputation, and validation for financial agents. Each agent can operate with a registered on-chain identity, while decision checkpoints create an auditable trail of actions, confidence, and reasoning. That gives the system a verifiable record instead of a black-box trading loop. Another major strength is risk-aware execution. NEXUS is built around policy controls rather than blind automation. It includes risk profiles, decision guards, a local kill-switch, recovery controls, and checkpointed runtime behavior so the agent can prioritize capital protection, drawdown awareness, and safer execution. This matches the hackathon’s emphasis on risk-adjusted performance, validation quality, and transparent agent behavior. NEXUS also stands out through its modular architecture. It combines PRISM-backed market and signal inputs, Groq-hosted Llama reasoning, algorithmic scoring modules, and a live Next.js dashboard for monitoring logs, checkpoints, system state, and controls in real time. In short, NEXUS shows that an autonomous trading agent can be configurable, transparent, risk-aware, and verifiable by design.

TrustTrade AI: Verifiable Autonomous Trading Agent

TrustTrade AI is a trust-minimized autonomous trading agent designed to operate safely in decentralized financial environments. The system combines AI-driven decision-making with on-chain verification to ensure that every trading action is transparent, explainable, and auditable. The agent analyzes real-time market data using intelligent strategies powered by a FastAPI and LangChain-based backend. It generates structured trade decisions including reasoning, confidence scores, and risk assessments. These decisions are converted into signed trade intents using EIP-712 and executed through a secure risk-controlled routing mechanism on the blockchain. To address the core challenge of trust in AI systems, TrustTrade AI integrates ERC-8004 registries for identity, reputation, and validation. Each action performed by the agent is recorded as a verifiable signal, allowing the system to build a measurable on-chain reputation based on performance, risk management, and validation quality rather than opaque outputs. Beyond execution, the platform introduces an advanced explainability layer that provides step-by-step reasoning, “why” and “why not” analysis, and confidence metrics for every trade. A replay engine allows users to trace decisions across time, while a strategy comparison and simulation engine demonstrates performance against alternative approaches. The system also includes dynamic risk intelligence, where the agent adapts its trading behavior based on drawdown, volatility, and historical outcomes. This ensures capital protection and responsible automation, moving beyond profit-only optimization. By combining AI intelligence, blockchain verification, and user-centric transparency, TrustTrade AI transforms trading agents from black-box systems into accountable financial entities. This project demonstrates a scalable foundation for deploying trustworthy autonomous agents capable of managing real capital in decentralized ecosystems.

FluxAgent

FluxAgent: Identity is the New Alpha The Black Box Problem & The ERC-8004 Solution Autonomous AI agents currently trade millions on-chain as complete black boxes, lacking cryptographic proof of reasoning and smart contract guardrails. FluxAgent solves this by introducing the first fully ERC-8004 compliant autonomous trading agent. We are transforming AI trading from opaque speculation into mathematically verifiable intelligence, ensuring every decision is signed, every trade is attested, and every risk is enforced on-chain. The Architecture: 2-Pass Critique & Immutable Guardrails Unlike standard agents that blindly trust their first LLM output, FluxAgent utilizes a Groq-accelerated 2-Pass Critique Architecture where the system acts as its own adversary, mathematically attacking its own logic to calculate a strict edge. Surviving trades are cryptographically signed via EIP-712 and routed through our RiskRouter.sol smart contract on Sepolia. This acts as an unbreakable circuit breaker, enforcing strict drawdown thresholds, position caps, and hourly limits. The LLM proposes, but the blockchain disposes. Traction & Deterministic Audit During live testnet trading, FluxAgent achieved an elite 100/100 Validation Score, maintaining a maximum drawdown of just 2 basis points across 15 approved trades. It successfully passed 12 out of 12 strict Phase-2 verification gates. Every 30-second tick generates a deterministic audit trail—reasoning traces, signatures, and on-chain fills are permanently logged and 100% reproducible. Built entirely solo, FluxAgent features fully deployed contracts and a live Vite/React console, proving that institutional-grade risk management can be completely autonomous.

Macro‑Sentry

Macro‑Sentry runs a full pipeline. First, it gathers macro and crypto signals. Then an LLM produces a strict JSON decision—BUY, SELL, or HOLD—along with a position size and reasoning. Next, the backend executes the trade via Kraken CLI. For a hackathon demo, we run safely in paper mode by default, but the same pipeline can run live if you configure Kraken CLI for a real account. For the ERC‑8004 trust layer, the agent can register an on-chain identity, produce EIP‑712 signatures for trade intents, and emit validation artifacts that create an auditable history of key decisions and actions. Those artifacts and reputation signals are displayed in the dashboard so judges can see not only what happened, but how it can be verified. Now I’ll show the demo. On the Dashboard, you can see performance metrics and the latest decision pulled from the backend. When I click Auto‑Trade Now, the system runs the complete loop: it fetches signals, the LLM outputs a decision, and we execute through Kraken CLI—paper mode in this demo. The result shows the action, an order id, the risk mode, and the reasoning. That trade is also logged to the Portfolio view so we can track returns and drawdowns over time. In on-chain mode, we additionally post a validation artifact linked to that trade, so it becomes verifiable on-chain and contributes to the agent’s reputation. What makes Macro‑Sentry competitive is that it’s not just a UI demo—it’s a deployable pipeline with safe defaults. You can run it with zero keys for a clean demo experience, and then switch to live execution and on-chain artifacts for real-world operation.

Apex - Self Certifying Yield Optimizer

APEX is an autonomous DeFi yield engine built for real execution, not just analysis. It combines four coordinated agents with on-chain reputation and live risk controls to turn fragmented yield data into disciplined trading actions. Scout gathers opportunities from DeFi protocols and market signals. Strategist filters, ranks, and sizes intents based on volatility, expected return, and portfolio constraints. Guardian acts as a circuit breaker, vetoing risky trades, enforcing drawdown limits, and protecting capital. Executor sends approved trades through Kraken and Surge, records outcomes, and posts evidence back to the reputation layer. The app streams every cycle in real time, tracks session PnL, open positions, fees, drawdown, win rate, and Sharpe-style metrics, and exposes a clear dashboard so users can see why the system traded or refused to trade. APEX is designed for transparency, accountability, and autonomy: the agents do not just suggest ideas, they operationalize decisions with verifiable logs, live telemetry, and a feedback loop that improves trust over time. The result is a trading interface that judges can inspect, users can understand, and the system can run continuously with measurable safety guarantees. It is built for judges, users, and builders who need a live, inspectable workflow, proof, and trust for every stakeholder!

ATLAS AI TRADING AGENT

ATLAS Swarm is a paradigm shift in how AI-driven automated trading operates. Most algorithmic trading bots operate in black boxes with no verifiable accountability. ATLAS changes this by requiring a multi-agent swarm to reach strict consensus before any trade is executed. We utilized Lang Graph to orchestrate specialized agents ranging from data accumulation nodes to sentiment analysts. Once the agent network identifies a profitable trade, it relies on the ERC-8004 identity registry deployed on Base Spolia. By securely tying the execution layer to a verifiable on-chain AI identity, every trade intent becomes cryptographically signed (EIP-712). Finally, for actual execution, we natively integrated the official Kraken CLI module into our workflow to handle programmatic paper trading. Our custom dashboard visualizes this entire flow: from the real-time swarm debate log to the final verified signature. This creates a fully accountable and transparent AI trading network.