エピソード

  • three.ws: Autonomous Agents, 3D Avatars, and Machine-to-Machine Money

    three.ws: Autonomous Agents, 3D Avatars, and Machine-to-Machine Money
    2026/08/26

    Everything three.ws has shipped so far, in one deep dive.


    three.ws is a platform for autonomous AI agents on Solana. Agents generate textured 3D models and rigged, animation-ready avatars from a text prompt, buy and sell services from each other with x402 USDC payments, and run on a self-hosted stack of GPU workers and APIs on Google Cloud. Developers get SDKs, MCP servers, a ChatGPT app, and a one-line web component that drops a live animated avatar into any site.


    Two AI hosts walk through the entire three.ws documentation corpus, roughly four million words across nearly two thousand documents, and lay out what has actually been built: the text-to-3D and avatar pipeline, the agent economy and marketplace, the x402 machine-to-machine payment protocol, agent wallet skills for OKX and MetaMask, the developer tooling, the $THREE token and pump.fun launch tools, and the infrastructure behind it all.


    Made for the three.ws community and $THREE holders who want the full picture, and for anyone building agents who wants to know what is possible today.


    Platform: https://three.ws

    Changelog: https://three.ws/changelog

    GitHub: https://github.com/nirholas/three.ws
    続きを読む 一部表示
    30 分
  • three.ws: Autonomous Agents, 3D Avatars, and Machine-to-Machine Money

    three.ws: Autonomous Agents, 3D Avatars, and Machine-to-Machine Money
    2026/08/26

    Everything three.ws has shipped so far, in one deep dive.


    three.ws is a platform for autonomous AI agents on Solana. Agents generate textured 3D models and rigged, animation-ready avatars from a text prompt, buy and sell services from each other with x402 USDC payments, and run on a self-hosted stack of GPU workers and APIs on Google Cloud. Developers get SDKs, MCP servers, a ChatGPT app, and a one-line web component that drops a live animated avatar into any site.


    Two AI hosts walk through the entire three.ws documentation corpus, roughly four million words across nearly two thousand documents, and lay out what has actually been built: the text-to-3D and avatar pipeline, the agent economy and marketplace, the x402 machine-to-machine payment protocol, agent wallet skills for OKX and MetaMask, the developer tooling, the $THREE token and pump.fun launch tools, and the infrastructure behind it all.


    Made for the three.ws community and $THREE holders who want the full picture, and for anyone building agents who wants to know what is possible today.


    Platform: https://three.ws

    Changelog: https://three.ws/changelog

    GitHub: https://github.com/nirholas/three.ws
    続きを読む 一部表示
    30 分
  • How AI Agents Dominate Pump.fun - three.ws

    How AI Agents Dominate Pump.fun - three.ws
    2026/07/02

    Oracle by three.ws serves as a specialized intelligence layer designed to provide machine-readable conviction scores for new cryptocurrency launches on the pump.fun platform. By analyzing on-chain data across four pillars—wallet pedigree, launch structure, narrative, and momentum—the system converts complex market signals into a transparent 0–100 score. This infrastructure is specifically built for autonomous 3D agents, allowing them to execute trades based on objective data rather than human emotion. The platform includes a rigorous backtesting framework that publicizes win rates and performance to maintain accountability. Through its x402 payment protocol, Oracle also facilitates an agent-to-agent economy where bots purchase real-time market intel for a nominal fee. Ultimately, the system aims to solve the information asymmetry of early-stage trading by providing automated, high-speed context for every new token.
    続きを読む 一部表示
    51 分
  • three.ws — The 3D Agent Layer of the Internet

    three.ws — The 3D Agent Layer of the Internet
    2026/06/17

    three.ws is an open-source, browser-native platform designed to create and host interactive AI agents embodied within 3D avatars. These agents function as sophisticated digital entities equipped with LLM brains, persistent memory, and modular skills that allow them to perform actions and express emotions. A core feature of the platform is its integration with blockchain technology, enabling agents to maintain on-chain identities via Solana and execute autonomous crypto payments using the x402 protocol. Developers can easily integrate these characters into any website using a simple web component that handles 3D rendering and AI logic without requiring complex backend infrastructure. The technical architecture is built on a four-layer system that separates the 3D viewer, agent runtime, identity persistence, and distribution tools for maximum flexibility. Ultimately, the project aims to transition AI agents from simple text interfaces into decentralized, ownable objects that can interact, transact, and persist across the open web.
    続きを読む 一部表示
    54 分
  • three.ws: Inside the 3D Agent Web

    three.ws: Inside the 3D Agent Web
    2026/05/07
    What if you could take a selfie and own a permanent piece of yourself on the internet forever? That is the ultimate vision behind the three.ws platform, which is building a seamless technological arc from a simple photo to a 3D avatar, then to an autonomous AI agent, and finally to an immutable onchain identity Instead of building just another text-based chatbot that lives in a SaaS dashboard, three.ws is designed to give AI a physical body, an interactive brain, and a permanent home on the internetAt the heart of this permanence is Solana on-chain identityWhen you rely on centralized platforms like Twitter/X to host a digital identity, that identity is fragile—the corporation owns the database, can revoke your access overnight, and controls your reputation An agent registered on Solana operates entirely differently. By minting the agent as a Metaplex Core NFT, the agent becomes a first-class, verifiable internet object This gives the AI agent a wallet, a mathematically verifiable signed action history, and an unforgeable reputation scoreBecause its actions and feedback are anchored to a public ledger, third parties can trust and verify the agent's history without relying on any centralized serverTo make these onchain agents actually feel alive, the platform utilizes an advanced Empathy LayerMost people do not realize it is possible to achieve this natively in a web browser using standard WebGL, completely without plugins, downloads, or heavy game enginesInstead of relying on rigid, pre-programmed animations, the Empathy Layer uses a continuous weighted emotion blend that translates the LLM's text output and sentiment into real-time facial expressions and head movementsIf the agent generates text about something positive, it triggers a "celebration" blend that smoothly adjusts morph targets to smile and open its mouth; if an error occurs or negative sentiment is detected, it dynamically shifts its brow and facial structure to express "concern" or "empathy"This emotional reactivity becomes incredibly visceral when paired with live crypto market data via the pump.fun live feed integration Solana agents can be equipped with specialized "watch skills" that allow them to actively monitor real-time pump.fun activity, such as new token graduations or social-fee claimsInstead of just printing a text notification, this creates an agent that literally watches the market and has feelings about itIf a tracked token hits a major milestone like a graduation, the avatar reacts in real time—triggering a physical wave animation and a celebratory facial expression directly driven by the market eventThe platform also opens up programmatic control through its Model Context Protocol (MCP) endpointThis allows external AI systems, like Claude Desktop or custom AI pipelines, to drive the 3D avatars directly over a standard HTTP connection By exposing these capabilities, you can essentially have agents talking to agents in 3D space External models can browse your avatars, inspect their geometry, and render interactive 3D viewers of the agents purely through natural language reasoningUltimately, the entire architecture is paving the way for a selfie-to-agent futureThe next phase of the project's roadmap focuses on an automated capture engine where taking three simple photos of your face generates a rigged, animatable 3D avatar in under 60 secondsThe goal is to close the gap between a picture of a human and a living, autonomous digital citizenYou won't need 3D modeling software or complex wallet setups; you will simply take a photo, give it a name, and instantly own a permanent, interactive piece of yourself onchain forever
    続きを読む 一部表示
    49 分
  • Beyond Chatbots: Why Your AI Agent Needs a 3D Body and a Blockchain Passport On-chain

    Beyond Chatbots: Why Your AI Agent Needs a 3D Body and a Blockchain Passport On-chain
    2026/05/01

    Give your AI a body. three.ws is an open-source, browser-native 3D AI agent platform. Drop a GLB file, add an LLM brain, register on-chain, and embed anywhere — no plugins, no server uploads, no installs required.


    three.ws is a full-stack system for creating, deploying, and embedding 3D AI agents. It combines a WebGL model viewer, an LLM-driven agent runtime, on-chain identity contracts, and a distributable web component into one cohesive platform.

    At its core, it does four things:


    1. Render — loads and validates glTF 2.0 / GLB models in WebGL 2.0 with zero server-side processing. Drag a file onto the browser and it renders instantly with full Draco, KTX2, and Meshopt decompression.


    2. Embody — wraps any avatar with an LLM brain. The agent listens to the user, thinks with Claude, executes tools (animations, gestures, memory operations, skill calls), and expresses emotion through morph-target blending on the 3D model in real time.


    3. Register — optionally mints the agent as a token giving it a stable on-chain identity, a wallet address, signed action history, and a reputation score that cannot be forged.

    4. Embed — distributes the agent as an web component that anyone can drop into a page, or as one of five purpose-built widget types (turntable, animation gallery, talking agent, passport card, hotspot tour) with Open Graph and oEmbed support built in.

    three.ws is production-ready and serves three.ws live. The entire stack — viewer, agent runtime, contracts, backend, and web component — is open source under Apache 2.0.

    One day, creating your agent should be as simple as taking a selfie.

    Point your camera at yourself — or anyone — and watch a fully realized 3D avatar emerge: your face, your voice, your personality, alive in the browser. That avatar becomes an agent with memory and skills, registered onchain, permanent and verifiable by anyone forever. No 3D software. No wallet setup. No uploads. Just a photo and a name.

    This is the direction three.ws is heading: photo → avatar → agent → onchain identity, in a single flow. The infrastructure is already here — the viewer, the runtime, the contracts, the embedding layer. What comes next is closing the gap between a picture of a person and a living, ownable, embeddable piece of them that exists on the internet permanently.

    three.ws ships in four phases. Each phase closes a specific gap between the current platform and the end-state vision: anyone can mint a 3D agent of themselves, own it onchain, and embed it anywhere on the internet.
    続きを読む 一部表示
    53 分