LiveScope per LiveScope

Field-by-field paste guide. The first one — Summary — looks fine, leave it. Walking the rest:

Description — paste:

LiveScope is a free, private reliability layer for AI chat. It scores every response from supported AI assistants in real time and surfaces the patterns most worth catching: unsupported claims, flattery openings, hidden hedges, fabricated specifics, and over-confident assertions.

What it does:

Adds a small reliability indicator to every AI response on supported platforms — ChatGPT, Claude, Google Gemini, Mistral Le Chat, and Microsoft Copilot.
Highlights the patterns that determined the score, so you learn the cues yourself, not just the verdict.
Tracks the trend across your last 50 responses with a sparkline in the popup.
Lets you flag false positives and false negatives directly from the overlay so the system gets calibrated to real-world failure modes.
Esc keyboard shortcut to dismiss the overlay when you disagree.
Localised into 47 languages with automatic locale detection.
Note for Firefox users: scoring on Firefox runs via a Cloudflare Worker we operate (the on-device WebLLM path requires a Chromium-only API not yet in Firefox). The Worker does not retain content beyond the request. No accounts, no analytics, no tracking pixels, no third-party SDKs.

Built by Everything Threads — an independent UK-based AI integrity company. Designed against EU AI Act compliance ahead of the August 2026 obligations. ICO registered: C1896585.