Anyone who has tried to move an AI agent across more than two blockchains knows the problem: the agent does not need ETH, SOL, AVAX, and TON to do its job, but it needs all of them to pay for the privilege of doing its job. Sydney-based Tychi Labs thinks that requirement is a bug worth killing.
Its answer is the Universal Gas Framework, or UGF, a developer-facing execution layer that lets users and AI agents transact across EVM and non-EVM chains without holding native gas tokens. Payment happens in stablecoins like USDC and EURC, or in a single native asset of the user's choosing, routed through a unified abstraction. "Pay with stablecoins like USDC, EURC, or native assets through a unified execution layer," the company writes in its own documentation [universalgasframework.com]. In practice, per a Tychi Labs post, a user can "pay gas once on Base with ETH and interact across multiple chains without needing native tokens everywhere" [Tychi Labs LinkedIn].
The bet
UGF is, at its core, a gas abstraction protocol pitched at developers and at the emerging class of autonomous on-chain agents. The wedge is narrow and specific: instead of asking a wallet, a script, or an LLM-driven agent to acquire and manage a native token on every chain it touches, UGF accepts a stablecoin or a single native asset and handles settlement underneath. The company's whitepaper describes "a breakdown of computation and storage costs, which are combined into a native gas fee estimate," which is then quoted to the user in their preferred denomination [universalgasframework.com].
The company's blog points at two adjacent surfaces it wants to serve. One is AI agents executing transactions autonomously [universalgasframework.com]. The other is x402-style payments, the HTTP-native micropayment pattern that Coinbase has been pushing as the rail for machine-to-machine commerce [universalgasframework.com]. Both share the same friction point: an agent that needs to make a one-shot payment on a chain it has never touched should not have to first go acquire that chain's native token through a centralized exchange.
Why it could be big
Gas abstraction is not a new idea. ERC-4337 account abstraction, paymasters, and meta-transaction relayers have been kicking around the EVM world for years. What is changing is the shape of the customer. Agentic systems, the ones doing the buying in an x402 world, are dreadful at managing fragmented token balances. They are excellent at calling a single API that accepts a stablecoin and returns a settled transaction. That is the curve UGF is positioned on.
The non-EVM angle matters too. Most existing gas-abstraction tooling is EVM-native. A framework that genuinely spans EVM and non-EVM chains, as UGF claims to [universalgasframework.com], collapses a real integration tax for any wallet, exchange, or agent platform that wants to be chain-agnostic by default. Tychi Labs is already shipping consumer infrastructure in the same neighborhood: the Tychi Wallet is live on the App Store [apps.apple.com], and in August Tychi Wallet announced a partnership with TaskOn aimed at Web3 community onboarding [cointrust.com, August 2025].
The team and traction
UGF is the work of co-founders Yash Singh and Rudr Rishi, both listed on the project's whitepaper and at Tychi Labs [universalgasframework.com] [LinkedIn]. The broader Tychi Labs team includes Nitin Kaushik, also tied to Tychi Limited on LinkedIn [LinkedIn]. The corporate footprint stretches across Sydney and a Tychi Limited entity registered in New Zealand [LinkedIn].
The public artifacts so far are a working documentation site, a whitepaper with a costed gas-estimation model, a series of explainer posts on AI agent execution and x402 payments, a shipped consumer wallet, and one named partnership with TaskOn for Web3 onboarding [cointrust.com, August 2025]. That is a coherent stack: a wallet that gives the team a distribution surface, a framework that gives developers a reason to integrate, and a partnership that points at user acquisition.
| Artifact | Status | Source |
|---|---|---|
| UGF documentation and whitepaper | Live | universalgasframework.com |
| Tychi Wallet (iOS) | Shipped | apps.apple.com |
| TaskOn partnership | Announced Aug 2025 | cointrust.com |
| x402 payments support | Documented | universalgasframework.com |
The honest counterfactual
The bear case is competitive density. Gas abstraction has become a crowded design space inside the EVM world, with paymaster infrastructure from established account-abstraction vendors and bridging-plus-gas products from several cross-chain teams. A new framework has to win on either developer ergonomics or on chain coverage that incumbents do not match. UGF's bull answer, on the evidence in its own documentation, is the second one: explicit support for non-EVM chains and an execution model designed from the start around stablecoin-denominated gas for AI agents rather than for human wallet users [universalgasframework.com]. Whether that differentiation holds depends on which non-EVM chains it actually covers at production quality, and which agent frameworks integrate first.
What to watch
Three things over the next twelve months will tell the story. First, integrations: a named agent framework, a named wallet beyond Tychi's own, or a named exchange picking up UGF would convert the thesis from interesting to durable. Second, the x402 question: if Coinbase's machine-payments pattern gains real volume, the gas-abstraction layer underneath it becomes infrastructure rather than a feature, and UGF has positioned itself early there [universalgasframework.com]. Third, a funding round. Tychi Labs has been shipping product without any disclosed investor on the cap table, and the next financing event will reveal which thesis the market is actually buying: the wallet, the framework, or both.
The deeper question for readers: when an AI agent needs to pay for a single API call on a chain it has never touched, who collects the toll on the on-ramp?