What happens if the key powering frustration runs out of credit?

It pauses rather than erroring or running silently: when the key funding frustration runs dry, Tessary marks the classifier no_credit, stops sending scoring calls, and skips every turn until the key is topped up or swapped for one with funds left.

An org’s own OpenRouter or TypeSafe key running dry pauses as no_credit; Tessary Cloud’s own key running dry pauses as a distinct platform_unavailable instead, since that failure is the deployment’s to fix, not the org’s. Enabling frustration doesn’t check credit at all: turning it on only fails when the lane has no key to run on, so an already-empty key can still be saved and switched on, and the classifier pauses the first time it tries to spend credit that isn’t there. Frustration is a hosted classifier call, not a per-trace LLM judge call, which is cheaper by design, but it still draws on the org’s own key.

The pause lifts on its own: saving a new key, topping up the old one, or pressing Retry on the rail all clear it immediately, and a background sweep rechecks a paused key on a fixed interval regardless. A turn skipped while paused stays unscored; nothing is scored retroactively.

sources

keep reading

More on this.

Two ways to run Tessary.

Tessary is an open-source agent reliability platform. Cloud and self-hosted run the same workflow on the OpenTelemetry traces your agent already emits.

Tessary Cloud

We host it for you. Send your first trace with nothing to deploy and no model key.

what's includedper organization
traces
10,000 per calendar month
stored trace data
1 GB
retention
30 days
model credit
$10, one-time, for triage and root-cause analysis
credit card
not required

Self-hosted Tessary

Run the open-source code on your own infrastructure with one command. Add your own model key for triage and root-cause analysis.

Self-host Tessary for me by following https://github.com/tessaryai/tessary/blob/main/setup.md

docker compose -f oci://docker.io/tessaryai/tessary:compose up -d -y