The Verified Toolkit · Live
The Verified Toolkit is a set of small, stdlib-only, independently-tested building blocks for agent and settlement systems. This is not a feature list — every button below runs a different real tool in your browser: an LLM-spend ledger with enforced caps, a rate limiter, a retry/circuit-breaker kit, an append-only event store, and a Markdown report builder. Each is deterministic, dependency-free, and shipped with its own test suite. Pure logic; nothing broadcast.
Each button runs a different real verified tool end to end. A tripped cap, a throttled request, or an open breaker is the tool working correctly — these are guarantees, not suggestions.
Raw tool output (spend ledger, rate-limit decisions, backoff schedule, event aggregation, rendered Markdown)
The Verified Toolkit — stdlib-only, independently-tested building blocks (spend ledger, rate limiter, retry/breaker, event store, report builder); nothing broadcast. Want the full toolkit (20+ self-contained tools) wired into your own stack? Talk to us →