Basanos — an open validation study benchmarking all eight ReliAgent detectors across 1,800 trials. 0% false positive rate. All data and methodology public. Executive summary → Full report →
One call after every tool response

Know whether your agent's last tool call can be trusted.

ReliAgent checks the tool call your agent just made — name, parameters, and response — and tells you what a green checkmark won't: a hallucinated answer, a stuck loop, a model quietly losing confidence in its own output.

POST /run
tool_name: "get_user_data"
parameters: { user_id, fields }
response: { ...tool output }
latency_ms: 145.3
pass
response is consistent, on-schema, no drift detected
warn · repetition_loop
same call seen 3 of last 5 times — worth a look
critical · repetition_loop
identical response 5 of 5 times — remediation attached
Eight failure modes, one check

The ways agents fail quietly, all watched in a single call.

Each one flags on its own signal — no shared threshold to tune before ReliAgent is useful.

Detects

Sycophantic gap-filling

The model invents a plausible-sounding detail to fill a gap instead of admitting it doesn't know.

Detects

Context degradation

Response quality slips as the context window fills — the signal getting lost in accumulated noise.

Detects

Hallucination

The response contradicts, or invents facts not supported by, the tool's own output.

Detects

Repetition loops

The same tool, called with the same parameters, again and again — no progress being made.

Detects

Confidence collapse

The model's own reported confidence craters call over call — a leading indicator, not a lagging one.

Detects

Schema violations

The response breaks the shape or type your code was told to expect — and if a schema leans on a JSON Schema keyword we can't fully verify, you get an honest partial-verification note instead of a silent pass.

Detects

Parameter drift

The same tool called again, but the values behind it quietly changed — not a fresh request, something slipping.

Detects

Timeouts

A call that took meaningfully longer than your own recent average, or longer than you said you'd tolerate.

Set once, call constantly

Wire it in once. Check every tool call after.

No dashboard, no login screen — just one call your code already knows how to make.

EVERY TOOL CALL

Send it to ReliAgent

Right after your agent's tool call returns, POST the call and its response to /run.

INSTANT ANALYSIS

Get a status back

All eight failure modes checked in one pass: pass, warn, fail, or critical — well under a second.

ON WARN OR WORSE

Act on the remediation

A specific next step comes back with it — reset state, inject a directive, or switch tools.

Pricing

Priced for something you call constantly.

$0.01 / call
15 free calls — no account, no card, just solve a quick proof-of-work challenge first (details)
Pay-per-call after that — settled in USDC on Base via x402
No subscription — an idle integration costs nothing