Shotoku releases local payment authorization layer for AI agents
Understand this story
Shotoku has released a local authorization layer for AI agents that decides to approve, hold, or deny payments before they occur based on budgets and rules. Policies are written in a single YAML file, supporting per-transaction maxAmount and a rolling 24-hour maxDailyAmount, with decisions recorded locally in data/decisions.jsonl. It does not custody funds or credentials, can intercept tool calls via an MCP wrapper mode, and provides Telegram approvals and a local dashboard. Shotoku says it runs entirely on the local machine, requiring no account, no cloud connection, and no telemetry. The product is currently in beta; Shotoku says it is already usable for real projects, but the API will still change; it is now open source and planned to develop into a commercial product.
AI overview based on source reports · Updated 13 hr ago
Coverage timeline
Follow the coverage to explore different perspectives.
- Hacker News AIShotoku – local authorization layer for AI agents
Shotoku is a local-first authorization layer that gives AI agents budgets, approval gates and a tamper-evident audit trail, enforcing spending decisions on your machine before any payment happens. It uses a YAML policy file with per-vendor maxAmount and maxDailyAmount limits, exposes a TypeScript authorize() function and an MCP server, and records every approved, denied or pending decision in a local JSONL file. Shotoku holds no funds or credentials and never signs or settles payments; it is in beta and open-source.
Activity for this story
- Current comparable activity
- 7
- Peak comparable activity
- 10Nov 6 01:00
- Change in the past 24 hours
- –
Trends compare the same fully observed participants, which may cover fewer participants than current activity. Move the pointer or tap the chart to inspect hourly activity; use the left and right arrow keys with a keyboard.