Herd
Collaborative terminal for contracts
Performances
RD Popularity Index
RD Growth Index
Comparison
Follow Updates
Follow Lists
People
About Herd
Herd is a collaborative terminal for the AI-agent era where people and AI agents jointly search, understand and compose on-chain contracts into executable on-chain or off-chain outcomes. It targets the pain point of 'contract visibility' - the same liquidity is wrapped repeatedly across protocols so neither users nor bots can see the real path and risk - by building contract indexing and an agent-collaboration layer that structures calls, parameters and risk warnings so agents act only after human confirmation. Its mechanism abstracts contract semantics into machine-readable interfaces, letting agents orchestrate multi-protocol interactions within constrained boundaries rather than merely producing text suggestions.
Herd has no token and has published no user or revenue data, with its product still in small-scale testing with a handful of partners, so whether it can build a sustainable fee model in the agent sector is its biggest uncertainty, and it must compete with general AI assistants and wallet-native agent features while improving the accuracy and coverage of its contract indexing.
Herd is building a collaborative terminal where people and AI agents search, understand and compose onchain contracts into executable onchain or offchain outputs.
On February 4, 2025 it announced a $1.8 million pre-seed led by Semantic Ventures with Archetype and Hardi Meybaum participating, when it had only a prototype and early partners.
It then focused on contract indexing and an agent collaboration layer, attacking contract visibility: the same liquidity is wrapped repeatedly across protocols, leaving users and bots unable to see the real path.
As of October 2026 Herd has no token and has published no user or revenue data, its product remains in small-scale testing with a handful of partners, and finding a durable fee model in the agent sector is its biggest uncertainty.
On June 22, 2026 Herd, built by a team led by former Dune Analytics data lead Andrew Hong, launched a new homepage that narrowed its offering to three parts: a human-readable block explorer, DoubleClick, which turns any token or vault into a balance-sheet graph, and Herd MCP, which plugs into Claude, Codex and Cursor. The explorer and MCP are open on Ethereum and Base, while DoubleClick remains in invite-only beta.
In July Herd walked through the roughly $6 million Summer.fi exploit on Bankless: an attacker hid a near-worthless Stream Finance-tainted token three layers deep inside more than 60 adapter positions across two vaults, and the curators never caught the dependency, while DoubleClick maps that chain of exposure directly. Steakhouse and Chaos Labs are already users; team size, revenue and funding have not been disclosed.
Herd's DoubleClick is still an invite-only beta, with full graph and monitoring capabilities limited to the enterprise plan; the next step is turning it into a subscription service for curators, comparing additions, changes and retirements in token and vault graphs period by period, tracking dependency shifts through monitors, and publishing research in the DoubleClick forum. On the explorer and MCP side, HAL adapters and transaction simulation are being extended to all registered EVM chains, with a Slack agent added.
Andrew Hong is founder and chief executive, with a BSc in Mathematics and Economics from New York University. He worked in credit risk, liquidity risk and CMBS analysis at Barclays, then as a data scientist at Mirror.xyz and Head of Developer Relations at Dune, founded Bytexplorers and created Crypto Data Bytes, and is now Research Advisor at Archetype. He launched Herd Labs as chief executive in 2024.
Edward Bramanti is co-founder and chief technology officer. He started on Dune's engineering and developer relations team, worked on on-chain data infrastructure and smart contracts, and co-founded Herd Labs with Andrew Hong in 2024 to lead protocol, contract and technical architecture.
The two previously ran engineering and developer relations respectively at Dune, making them the most closely paired team members.