Dante AI
Turn more visitors into customers with your own AI chatbot – no tech skills needed. Dante AI helps your website work smarter by adding a helpful, human-sounding AI chatbot to your WordPress site. It can answer questions, guide visitors, and help you generate more leads – automatically. Whether you’re running a blog, a local business, or an online shop, Dante AI gives your visitors instant support – so they stay longer, trust you more, and convert faster. Key Features Quick Setup – Add your AI chatbot in minutes Create AI Chatbots Easily – Set up and manage your chatbot directly from your WordPress dashboard One-Click Training – Instantly train your AI chatbot using your WordPress website’s existing content Customize the Look and Personality – Match your AI chatbot to your brand with simple appearance settings Track Performance – See how your AI chatbot is helping with easy-to-read stats Mobile Friendly – Works beautifully on phones, tablets, and desktops How It Works Sign up for a free Dante AI account Create your chatbot (no coding needed) Copy your API key Connect it to your WordPress site You’re live! Start engaging your visitors automatically Data and Privacy We take your privacy seriously – and that of your visitors. Here’s what you need to know: You stay in control – Chat data belongs to you and is visible only in your Dante AI dashboard No personal info is collected – Unless a user voluntarily types it in during a conversation Secure and private – Conversations are processed and stored securely on Dante AI servers Easy to turn off – You can disable the chatbot any time from your WordPress admin panel For more details, see our Privacy Policy and Terms of Service
Npcink AI Client Adapter
Npcink AI Client Adapter connects local AI clients, such as OpenClaw-compatible tools, to a WordPress site through a focused REST adapter. It is designed for sites that use Npcink Governance Core for reviewable AI operations and WordPress Abilities API providers, such as Npcink Abilities Toolkit, for concrete WordPress capabilities. The Adapter gives clients one stable WordPress REST surface for checking readiness, reading governance guidance, running approved read abilities, creating governed proposals, and executing supported writes only after Core approval and commit-preflight. The default connection path uses signed key-pair pairing so the AI client does not need to receive a WordPress Application Password. A WordPress Application Password fallback is available for clients that already have a dedicated secret field or credential vault. Adapter is intentionally thin. It does not define abilities, store approval truth, run workflow queues, act as an MCP server, manage model providers, store prompts, or execute arbitrary final write mutations. Governance remains in Npcink Governance Core. Ability definitions and callbacks remain in Npcink Abilities Toolkit or another WordPress Abilities API provider. Adapter exposes a machine-readable client_policy on health, help, and the connection manifest so compatible clients can consume explicit route, output, sensitive-read, and write-flow boundaries. The local CLI also redacts profile paths, key ids, signing headers, tokens, passwords, and secrets from output. Current governed execution support covers explicit Adapter profiles after Core approval, including npcink-abilities-toolkit/trash-post, npcink-abilities-toolkit/create-draft, npcink-abilities-toolkit/update-post, npcink-abilities-toolkit/update-post-blocks, npcink-abilities-toolkit/set-post-terms, npcink-abilities-toolkit/reply-comment, and npcink-abilities-toolkit/approve-comment. Key features: Connect OpenClaw-compatible and similar local AI clients to WordPress through one Adapter REST namespace. Check Adapter readiness and dependency status before a client starts work. Expose a machine-readable client_policy so clients can understand route, read, write, and sensitive-data boundaries. Route approved direct-read requests through the WordPress Abilities API. Forward governed write requests to Npcink Governance Core proposal and commit-preflight endpoints. Support a user-triggered approve-and-execute path for explicit, allowlisted execution profiles after Core approval. Prefer signed key-pair pairing for local clients, with an Application Password fallback when appropriate. Keep channel, governance, ability, cloud, and model-provider responsibilities separate.