AnswerRails™

• It won't embarrass you. Every grounded answer carries real source citations. When your content doesn't hold the answer, it admits it — no fabrication, no bluffing, no confident nonsense in your name.
• No middleman on your data. Conversations and chat logs live in your own WordPress database, and your knowledge syncs to your own AI-provider account — AnswerRails never sees, stores or meters your data, and there's no third-party platform in between. Consent, retention and privacy controls are built in, not bolted on — and provable, with server-side consent receipts and a privacy audit trail.
• We read the provider terms so you don't have to. We dropped Google Gemini when its API terms didn't fit an embedded public-site chat. We won't ship a provider whose terms put your site at risk — that diligence is part of the product.
• It tells you the truth about itself. A health dashboard shows in plain language what's answering, what's failing and why — and warns you before a model you rely on is retired. No silent degradation.
• Trust is not an upsell. Citations, honest answers, privacy machinery and health visibility are free on every site. Pro adds the leverage layer: provider choice, deep grounding, product & URL knowledge sources, analytics, lead and idea capture, and human escalation.

SMEs running WordPress who want fewer repetitive emails, trustworthy answers, and faster resolutions that keep customers happy.

• Themeable chat widget — match your brand colour, with automatic dark mode and contrast-safe text. Visitors can move and resize the window on desktop; phones get a clean full-screen chat.
• Accessible to every visitor — built and tested to WCAG 2.2 AA: screen-reader announcements, full keyboard operation, visible focus, reduced-motion support, and contrast-checked theming. Verified on real assistive technology (NVDA and iPhone VoiceOver) — read the accessibility statement.
• Knowledge from your pages, posts and product data — answers cite their sources, and items stay in step with the content they came from. If a source page disappears, the assistant stops using it — with a clear status in your dashboard — rather than quietly serving stale text.
• Anti-fabrication safety rails: it admits what it doesn't know rather than guessing.
• Privacy-aware logging with consent and retention controls — consent is recorded as server-side receipts (including the exact notice agreed to), and visitors can withdraw it from inside the chat as easily as they gave it.
• Accountability built in: GDPR data-request tooling on every plan, plus a privacy audit trail recording legal holds, data-subject requests and blocked operations. Chat logs show which safeguard path produced every answer — and even a failed send during a provider outage leaves a record of exactly what the visitor saw.
• Visitor-input moderation on every plan — abusive messages are deflected before they ever reach your AI provider account, self-harm signals get a supportive reply with helpline signposting, and repeated abuse ends the session. Pro can tune the responses; the protection itself has no off-switch.
• Spam and abuse protection built in — rate limiting plus optional Google reCAPTCHA v3 with a tunable strictness threshold. Privacy-conscious: Google's script only loads for visitors who actually open the chat.
• Your choice of AI provider — OpenAI, Mistral, or Anthropic Claude — switchable in one click (Pro; the free tier runs on Mistral's zero-cost API tier).
• Human handoff (Pro): configurable escalation triggers for words like complaint and refund and phrases like speak to someone, with a structured escalation email so nothing gets lost.
• Conversation analytics that show answer quality and the gaps in your knowledge base — plus idea capture that turns real visitor questions into free market research (Pro).

Feature	Free	Pro
Chat widget + on-site styling	✓	✓
Answers grounded in your content, with source citations	✓	✓
Anti-fabrication safety rails & structured, policy-checked answers	✓	✓
System health dashboard & diagnostics	✓	✓
Privacy-safe defaults, consent receipts & retention controls	✓	✓
GDPR data-request tooling & privacy audit trail	✓	✓
Visitor-input moderation (abuse & self-harm safeguards)	✓	✓
Spam protection (rate limiting + reCAPTCHA v3)	✓	✓
AI provider & premium-model choice	Mistral free tier	OpenAI · Claude · Mistral
Grounding depth (knowledge passages per answer)	Essential — top 5	Deep — adjustable dial
Knowledge sources	Pages, posts, documents, JSON	+ WooCommerce products & external URLs
Conversation analytics & knowledge-gap insights	—	✓
Sales-lead detection & idea capture	—	✓
Human-handoff escalations with email transcripts	—	✓
Advanced privacy controls (storage mode, auto-clear, identified logging)	—	✓

The trust layer — citations, honest answers, health visibility — is free on every site. Upgrade to Pro for the leverage layer: your choice of providers and models, deep grounding, product & URL knowledge sources, analytics, capture, and human escalation.