Skip to content

Releases: stevenknowswhy/article50-voice-agent-proof

v1.2.0 — fictional voice qualification demo

Choose a tag to compare

@stevenknowswhy stevenknowswhy released this 01 Aug 14:13

Adds a listenable local text-to-speech rendering of the fictional qualification flow and its exact public transcript. The demo identifies the AI agent, keeps advice and data-capture boundaries explicit, and requires human review before any private checkout. It is not a recording of a phone call and does not prove a number, carrier connection, inbound routing, outbound permission, latency, or production telephony. The public repository contains no phone number, provider credential, lead, or checkout URL. Verified by 17 tests plus Ruff in GitHub Actions run 30703297264.

v1.1.0 — permission-gated call lifecycle simulator

Choose a tag to compare

@stevenknowswhy stevenknowswhy released this 01 Aug 13:23

Adds a deterministic local simulator for inbound disclosure, permission-gated outbound calls, recording consent, opt-out handling, and data-minimized capture. The simulator creates no live telephony resource and is not evidence of carrier connectivity. Verified by 15 tests, Ruff, GitHub Actions, and isolated source and wheel builds.

v1.0.0 — inbound voice-agent proof

Choose a tag to compare

@stevenknowswhy stevenknowswhy released this 01 Aug 12:08

Secrets-free reference implementation for inbound AI voice qualification. Includes explicit AI disclosure, consent-before-capture, counsel-owned legal decisions, opt-out handling, sensitive-data and private-checkout exclusions, raw-value logging suppression, LiveKit and Vapi adapters, six guardrail tests, and strict linting. This release does not create or expose a live phone number.