SpendNod is the authorization layer for AI agent transactions. Your agent connects via MCP, checks in before every purchase, and waits for your approval on anything above your threshold. No code required.
Works with every MCP-compatible platform
Agents can book flights, make purchases, and transfer funds. But most operate with zero human oversight between the decision and the transaction.
AI agents can trigger unlimited transactions with no caps, no category restrictions, and no velocity limits.
When something goes wrong, there's no record of what the agent did, why it did it, or who authorized it.
The EU AI Act requires auditable proof that AI systems operated within authorized parameters. Most agent frameworks have no answer for this.
Add SpendNod to your agent's MCP config. Set your rules. The engine decides in milliseconds.
Add the SpendNod MCP URL to your agent's config file. Works with Claude, ChatGPT, Cursor, and any MCP-compatible platform. No SDK install, no code changes.
Configure spending thresholds, vendor blocklists, daily caps, and category restrictions from the dashboard. Start with a template - Conservative, Moderate, or Permissive - or build custom rules.
Before every transaction, your agent calls SpendNod. The rule engine evaluates in milliseconds and returns one of three outcomes:
Auto-approved Under your threshold - instant, no human needed.
Pending review Above your threshold - you approve with one tap.
Denied Blocked vendor or limit hit - instant rejection.
Every decision is logged - who approved it, which rule triggered, timestamp, full context. Exportable and compliance-ready.
Built for developers who want zero-friction integration and humans who want real control.
Add one URL to your agent's config. SpendNod becomes a native tool - your agent automatically calls authorize before every transaction. No code required.
Per-transaction limits, daily/monthly caps, vendor blocklists, category restrictions, velocity limits. Start with preset templates or build custom rules.
See pending requests with full context - agent name, vendor, amount, description, countdown timer. Approve or deny from your phone, tablet, or desktop.
New agents require human approval for everything. You loosen controls as trust builds. Safety first, always.
Works with Claude, ChatGPT, LangChain, CrewAI, AutoGen, or custom agents. Works with any payment rail. SpendNod sits above all of them.
Every request, every decision, every rule change - timestamped and logged. Know exactly what happened and who authorized it.
Lock in doubled limits forever. Prices go up after the first 500 signups.
Launch pricing locks in for the lifetime of your account. Every plan includes MCP server, REST API, Python SDK, rule engine, dashboard, and audit logging. You only pay for scale.
Add one URL. Set your rules. Keep your humans in the loop.