Sovereign AI Workspace
YveChat connects a hosted workspace to AI models running on your computer, while your models, conversations, images, and compute remain on your machine.
Local AI models and runtimes
Backed by open cryptographic records and digital signatures.
Achievement counters use a separate non-tradable evidence contract on Base. $YVE remains pre-TGE and is planned for BNB Chain.
A static workspace served from app.yvechat.xyz.
A local connector, paired by a six-digit code.
Ollama and ComfyUI running on your hardware.
Conversations and credentials stay on your machine.
YveChat runs local models through a component on your computer. Six things get you started.
Windows, macOS, or Linux with enough memory for the models you choose.
The connector that links the hosted workspace to your machine. Free to download and run; currently an unsigned developer preview, so your operating system will warn you before it starts.
The developer preview runs on Node.js. Install it before the Engine, or the launcher will stop and tell you to.
Ollama is the primary supported runtime. vLLM and OpenAI-compatible endpoints have experimental support.
Pull a model with Ollama before you begin. Availability depends on what you install.
A current browser that can reach your local runtime once you grant permission. Chrome, Edge, Brave, and Firefox are verified. Safari is not, and can open YveChat directly at 127.0.0.1:3210 instead.
Web search runs today through the local tool gateway and returns clickable sources. Image generation renders locally through ComfyUI. Page and PDF reading are experimental in the same gateway. Local document retrieval and code execution sit on the roadmap.
Download the Engine and run it on your computer so the workspace can reach your machine.
Download YveChat EngineSet up Ollama and pull at least one model. Larger models need more memory and stronger hardware.
Open the workspace, then confirm the six-digit code on your machine to pair.
YveChat discovers your installed Ollama models after the Engine connects. Pick one and start.
YveChat Engine required
Each era hands another layer of control to the user.
Custody
Local models, local data, and local pairing on your own machine.
EngineNow
Local inference through an explicitly paired connection. Signed installers are next.
Gateway
Web search is available. Page and PDF reading are Experimental ahead of full chat integration. Sandboxed execution and MCP remain Planned.
Network
An opt-in compute network for shared capacity.
Sovereignty
Governance over protocol parameters and independent oversight.
Utility arrives in layers across the roadmap, tied to the capabilities each era introduces. Every $YVE utility is planned.
$YVE remains pre-TGE and will be deployed on BNB Chain. No contract address has been published and no sale is open today. Treat any claimed address or offer as unverified until it appears on Verified Links.
Staking unlocks the scaling tier.
Meters tool calls on hosted providers.
Settles listings and revenue share.
Pays operators in the compute network.
Weight over protocol and treasury.
Dan sees AI becoming the record of how we think. Research, half-formed ideas, decisions, questions, and years of conversation now live inside AI workspaces. He believes ownership of that intelligence will be one of the defining questions of AI's next phase, and it is the reason YveChat keeps yours in your custody.
Alvis is obsessed with local AI: model infrastructure, self-hosted data, provider-neutral architecture, and agent tooling. He believes the next phase of AI belongs to whoever controls the infrastructure around intelligence. YveChat is his answer, a workspace where you keep your models, conversations, compute, credentials, and identity. Your models. Your machine. Your workspace.
Follow the setup to run local models today. Read the docs for the full architecture.
Built for people ready to run their own AI.