Zero1 Labs (𝟘,𝟷)

AI depends on data. Blockchains depend on trust. To combine them, privacy has to be solved. Cypher is Zero1 Labs’ FHE–EVM chain, designed for AI and LLM applications. It performs computations directly on encrypted data, ensuring inputs stay private while results remain correct. Built for confidential workloads, Cypher preserves scalability, enforces compliance standards, and enables developers to deploy decentralized AI applications without exposing sensitive information. Learn more: cypher.z1labs.ai

The full Zero1 Labs codebase is open-source. Our GitHub includes encrypted compute primitives, on-chain coordination tools, MCP integrations, and governance modules. Explore the repo → github.com/z1labs

Modern AI has a data ownership problem. Training datasets are often scraped without consent, hard to audit, and controlled by centralized platforms. Z1 is building infrastructure for decentralized AI with encrypted data pipelines and verifiable provenance.

AI can now compute on data it never sees. Cypher is the FHE-EVM chain, purpose-built for confidential AI. Encrypted inputs, verifiable outputs, full privacy by default. Deploy decentralized AI without ever exposing sensitive data.

Turn AI chats into onchain NFTs. Chat-NFT pins each session to IPFS, hashes context + model metadata via MCP, and mints an ERC-721 to your wallet. Explore → mcp.z1labs.ai

Cypher Testnet Milestone: February 2026 83.9M transactions on Cypher testnet, processed entirely under FHE. Data stays encrypted through execution. • 6.89M addresses • 122.1K encrypted TXs/day • 12.7M DEAI transfers • 35.6M blocks produced • 257.1 GB encrypted state processed • 193.8 GB storage used • 65,251 uploads in a single day Cypher is the EVM-ready network for encrypted computation, bringing private smart contract execution to real-world scale.

The MCP ZK Integrity Layer is live on Z1 GitHub. It adds zero-knowledge verification to the Model Context Protocol. Agents can generate proofs that: • a computation was executed correctly • context wasn’t altered • outputs match valid inputs All without revealing private data. Verifiable integrity, built directly into MCP execution. github.com/z1labs/mcp-zk-…

Built natively for AI apps. Engineered for encrypted execution. Cypher integrates FHE, concurrency, and composable architecture to support scalable, privacy-preserving computation on-chain. Execution remains encrypted at every stage.

Privacy is not about hiding data at rest. It’s about eliminating the moment of exposure. In most systems: encrypted → decrypted → computed → exposed. With Cypher: encrypted → computed → encrypted. Execution never transitions into plaintext. That changes what autonomous systems can safely do on-chain.

Fully Homomorphic Encryption (FHE) changes where privacy lives in a system. In most architectures today: 1. Data is encrypted while stored or transmitted. 2. It is decrypted before computation. 3. Execution happens in plaintext. That execution step is the exposure point. FHE removes that requirement. Computation is performed directly on encrypted data. The system never needs access to the raw inputs to produce a valid result. The output remains encrypted and can only be decrypted by the key holder. This makes confidential execution possible: - Smart contract logic without exposing inputs - AI inference without revealing model context or user data - Agents operating without leaking internal state Cypher is Z1’s implementation of this model, applying FHE at the execution layer so state does not need to be revealed for computation to occur.

MCP Context Badge Claim Module github.com/z1labs/mcp-nft… Implementation for minting ERC-1155 badges only when specific AI context conditions are met. Each claim is checked against the model used, a referenced context CID, and the claimant’s wallet. Proofs can come from an agent signature or an offchain verifier before minting proceeds. • Context + model checks (contextCID, modelHash) • Wallet-signed claims • Agent-signed or offchain proof support • ERC-1155 mint with context-linked metadata • Works via frontend or scripts Useful for cases where badges or credentials need to be tied to real interaction context.

Mintable AI Agents on Cypher, the first standard for turning trained models into verifiable digital assets. Each agent is an NFT bound to its SHA-256 model hash, with weights stored on IPFS or Arweave. Immutable. Verifiable. Optionally encrypted with FHE so your model stays private while proving it's real. Developers can monetize, license, and build an on-chain portfolio. Collectors can own functional AI with provable authenticity and performance history. → mcp.z1labs.ai

Most privacy systems protect data, not computation. Cypher uses FHE so execution happens on encrypted state. No point where data needs to be revealed.

More activity ≠ less encryption. Cypher's testnet handles growing load while keeping FHE guarantees intact. More transactions, more state. Data and computation stay encrypted end-to-end. Learn more about Cypher: cypher.z1labs.ai

Cypher Testnet Milestone: January 2026 81.5M transactions on Cypher testnet, processed entirely under FHE. Data stays encrypted through execution.​​​​​​​​​​​​​​​​ - 6.89M addresses - 120.5K encrypted TXs/day - 10.2M DEAI transfers - 35.3M blocks produced - 253.1 GB encrypted state processed - 190.2 GB storage used - 64K uploads in a single day EVM-compatible, scaling toward mainnet conditions.

All Zero1 Labs repositories are public on GitHub. Cypher, tooling, and supporting infrastructure are available to explore, run, and build on. → github.com/z1labs

Autonomous systems need private execution. When everything is public, agents expose how they plan and coordinate. That limits what they can safely do over time. Cypher provides a private execution layer where agents can run without exposing internal state, while still producing on-chain results.

As we start 2026, it’s worth looking back at what Zero1 Labs delivered in 2025. Throughout the year, the Cypher testnet processed millions of on-chain actions, validating system stability under sustained usage and continuous upgrades. One of the most important releases was persistent AI agents. Agents now maintain identity, execution history, and context across sessions, enabling long-running, stateful workflows. We introduced agent credentialing and reputation tracking, allowing completed tasks and behavior to be recorded in a structured, verifiable way rather than relying on off-chain trust. Cypher underwent multiple phased testnet upgrades, each expanding execution logic, coordination rules, and system monitoring without interrupting network activity. To support transparency, we launched a public testnet dashboard, exposing live network activity, task completion metrics, and system usage for builders and contributors. We also expanded agent-to-agent coordination, enabling collaborative and multi-step execution flows instead of isolated, single-agent interactions. By the end of 2025, Cypher had evolved into a clear foundation for privacy-preserving, EVM-compatible execution. Looking ahead to 2026 milestones.

2026 begins. Thank you to the community and partners advancing privacy-preserving infrastructure with Zero1 Labs. Onward into another year of building.

Cypher Testnet Milestone: December 2025 Since launch, the Cypher Testnet has processed over 80.4 million transactions, reflecting sustained network activity and growing adoption of encrypted computation. • 6.82M total users • 118.9K encrypted daily transactions • 35.1M blocks produced • 9.94M DEAI transfers • 241.6 GB encrypted data processed • 178.9 GB total storage used • 63,480 uploads in a single day Cypher is building EVM-compatible infrastructure for encrypted computation, enabling privacy-preserving execution at production scale.

Merry Christmas from Zero1 Labs. Thank you for being part of our story this year. More to come in 2026.