What AgentLedger's proxy actually costs you in latency, measured live — not a claimed number.
Time spent computing whether a call would exceed budget, before forwarding to the provider. This is the AgentLedger tax — it does not include the provider's own response time, which AgentLedger does not control.
From 6 measured calls on this instance (resets on deploy — see /status for current uptime).
If the upstream provider is unreachable, the proxy returns 502
upstream_unreachable to the caller rather than retrying — a retried
call against a provider that already received the request risks a double
charge, so this product never retries automatically.
/status reports uptime and version for this instance. /health is the liveness check. Uptime is per-instance: a deploy restarts the process.
AgentLedger — not affiliated with agentledger.co or other same-named projects · start in 5 minutes