SpendVeto: The Buyer-Side Governance Layer for AI Agents

Introduction

SpendVeto is an open-source tool designed to control how AI agents spend money. As AI agents gain the ability to hold digital wallets and pay for services like APIs and data on their own, they create a risk of overspending if left unchecked. SpendVeto acts as a safety layer between the agent and the payment system. It ensures that every transaction follows strict rules, stays within budget, and gets human approval when needed before any money changes hands. This tool helps teams manage the financial risks of autonomous software.

Benefits

SpendVeto offers several key advantages for organizations using AI agents:

• Strict Budget Control:The system enforces per-call limits and hourly budgets. If an agent tries to spend more than allowed, the transaction pauses for human review. It also supports nested budgets where parent accounts grant limited funds to child accounts, ensuring no single part of the system can overspend.
• Human Oversight:When spending goes above a set threshold, the system stops and waits for a real person to approve or deny the action. If no human responds within a set time, the transaction fails automatically. This prevents costs from accumulating while a team member is away from their desk.
• Runaway Detection:The tool can detect if an agent is trying to make payments too quickly, which often signals a bug or a security breach. It can instantly freeze the agent to stop the bleeding. Operators can also manually freeze any wallet from a dashboard at any time.
• Secure Architecture:In its enforcement mode, AI agents never hold the private keys needed to sign payments. Instead, they send spending requests, and a secure proxy wallet signs them only after all safety checks pass. This ensures a rogue agent cannot bypass security rules.
• Multi-Chain Support:SpendVeto works across 15 different blockchains. It ensures that an authorization made for one blockchain cannot be used to pay on another, preventing unauthorized cross-chain transactions.

Use Cases

SpendVeto is ideal for teams deploying AI agents that need to make purchases autonomously:

• Autonomous Agents:Companies using AI agents to buy software licenses, pay for cloud computing resources, or purchase data feeds can use SpendVeto to set spending limits and require approvals for large purchases.
• Developer Teams:Developers can integrate SpendVeto using its command-line tools or software development kits. This allows them to build agents that can self-correct if they try to spend money outside of allowed rules.
• Enterprise Finance:Organizations looking to manage agentic commerce can use SpendVeto to create auditable ledgers. Every transaction is signed and recorded, making it easy to track exactly where money went.
• Testing and Development:Since the tool is open-source and free to self-host, teams can test their AI agents in a safe environment without risking real money. They can run built-in tests to verify that their agents follow spending policies correctly.

Pricing

SpendVeto is currently free to use for self-hosting. The core product is open-source and licensed under Apache-2.0, meaning teams can download, install, and run it on their own servers without paying a fee. There is no vendor lock-in, and users do not need to provide private keys to test the system. A managed hosting platform with features like single sign-on and organizational accounts is available on a waitlist for those who prefer a hosted solution.

Vibes

The project has received positive attention for filling a critical gap in the AI security market. By focusing on the buyer-side of AI spending, SpendVeto addresses a problem that existing expense platforms and seller-side tools do not solve. The community appreciates its transparency, as every feature is backed by a real, running test suite. Users value the ability to verify signed receipts and export full ledgers for accounting purposes. The open-source nature of the project has built trust, allowing developers to inspect the code and ensure it meets their security needs.

Additional Information

SpendVeto was created to address the growing volume of stablecoin transfers and the projected trillions of dollars in agentic commerce by 2030. The project is built to be developer-first, offering native adapters for popular frameworks like LangChain and OpenAI Agents. It supports multiple signature families to ensure compatibility across different blockchain ecosystems. The team emphasizes that while payment rails handle the mechanics of paying, SpendVeto solves the critical question of how to stop agents from overspending before the money moves.