KinSvarmo

KinSvarmo is a marketplace and workspace for private AI agents represented as 0G iNFTs. A user creates or selects an agent, uploads a dataset, authorizes agent usage on 0G Galileo testnet, and starts a job. The backend coordinates the job through Gensyn AXL nodes using planner, analyzer, critic, and reporter roles. 0G Compute is used for model inference when configured, and the result records provider, model, chat ID, and output metadata. KeeperHub records the workflow execution, node statuses, and logs for each run. The UI shows the agent, job state, AXL message flow, KeeperHub execution state, 0G compute proof, and final structured result.

KinSvarmo uses a Next.js frontend on Vercel, a Node/Fastify API on Railway, 0G Galileo testnet contracts, 0G Storage, 0G Compute, KeeperHub, and Gensyn AXL. The UI creates agents, uploads datasets, signs 0G usage transactions, starts jobs, and reads job state from the API. The API stores runtime job state in memory, calls KeeperHub to create a workflow execution record, then sends job messages through AXL nodes. The AXL path is API, planner, analyzer, critic, reporter, then API. The Railway service runs the API plus real AXL node binaries; a TCP proxy allows an external reporter node to connect from another terminal or machine. 0G Compute performs the model call and returns provider, model, usage, chat ID, and output data. KeeperHub adds an execution trace and logs around the workflow.