Konnex Launches Onboarding Phase For Developers And Miners Ahead Of On-Chain Testnet

Konnex, the developer behind the decentralized market protocol for autonomous systems and robotics, launched “Konnex: Ingest,” marking the initial phase of developer and miner onboarding in preparation for the platform’s on-chain testnet scheduled later this year. 

The Ingest phase is designed to integrate Vision Language Action (VLA) models and Language Behavioral Models (LBMs) into the Konnex ecosystem within a controlled, off-chain environment, focusing on validating model compatibility, telemetry standards, and execution workflows before introducing economic or on-chain coordination mechanisms.

During this stage, contributors submit models to perform a predefined set of robotic task scenarios. Each execution produces structured telemetry and video outputs that can be monitored directly through a browser interface. Human evaluators then review these outputs, providing feedback on task success, safety, and behavioral quality. This feedback is incorporated as Reinforcement Learning from Human Feedback (RLHF) signals, allowing models to be iteratively refined based on actual task performance rather than synthetic or simulated labels.

The Ingest phase encompasses model submission and formatting validation, off-chain task execution and replay, telemetry collection for downstream verification, human-in-the-loop evaluation, and preliminary performance benchmarking across fixed scenarios. Operating off-chain enables fast iteration, debugging, and refinement without introducing economic risk or the need for validator incentives.

Konnex Outlines Roadmap For Onboarding, Runtime Testing, And Preflight Simulation 

Over the following weeks, Konnex plans to onboard additional contributors and expand the range of supported model formats. Data and feedback collected during Ingest will feed into “Konnex: Runtime Zero (R0),” where models will execute under full runtime constraints. Subsequent development includes “Konnex: Preflight,” a three-dimensional simulation environment designed for multi-agent coordination, validator replay, and pre-economic stress testing ahead of full on-chain deployment.

Documentation, miner calculations, and validator requirements will be released progressively as each component stabilizes. This phased strategy is intended to ensure that Konnex’s eventual on-chain operations are grounded in empirically observed behavior, verified human feedback, and repeatable execution, rather than theoretical assumptions.

Konnex is a Web3-native, permissionless decentralized marketplace built on the Solana blockchain, designed to enable autonomous robots to identify work opportunities, engage with AI service providers, exchange intelligence, and settle completed physical tasks on-chain through smart contracts and stablecoins, with outcomes verified via a Proof-of-Physical-Work system and coordinated by a decentralized validator network.

Earlier this month, the platform closed a $15 million strategic funding round to advance its on-chain physical economy initiatives. The investment, led by Cogitent Ventures and supported by Liquid Capital, Leland Ventures, Covey Network, Ventures M77, and Block Maven LLC, is intended to support the development of infrastructure for scheduling, verifying, and compensating autonomous robotic work on-chain, with the broader goal of integrating real-world labor into blockchain networks.

The post Konnex Launches Onboarding Phase For Developers And Miners Ahead Of On-Chain Testnet appeared first on Metaverse Post.