Inspect the evidence before accepting an action.

Truquence compares supplied n8n execution JSON or Make scenario logs against explicit expectations and produces a local PASS, FAIL or UNKNOWN report. It also retains the synthetic three-checker Action Contract merger.

FAIL and UNKNOWN block acceptance. Reports are unsigned. The tool does not execute actions or establish the authenticity of supplied evidence.

Install and call

python3 -m pip install .
truquence init --workflow workflow.json --out contract.json
# Fill independent expectations, then:
truquence accept --format n8n --contract contract.json --execution execution.json --out acceptance

Run from a source checkout with Python 3.10+. Installation may fetch the build tool; runtime has no remote dependency. The repository README links a real local n8n walkthrough with a fresh workflow, no packaged samples and no account.

Read the contract

Workflow formats · Local n8n walkthrough

Input schema · Evidence and output reference · Agent reference · Loopback CSV API

Measured limits

Read the historical synthetic comparison. Under its conditions, an interface around the same core added latency and did not establish an algorithmic or human-effort advantage. It does not benchmark the current Action Contract implementation.