Yes. SleekAI reads any post type in your WordPress install, so a single bot can answer from /docs, /sdk, /recipes, /status, and /changelog at once. For large portals, scope with multibot, a docs bot on /docs and a recipes bot on /recipes, each with its own system prompt, so the answer style matches the page the developer is on.

If your status feed lives inside WordPress (a CPT or block-based posts), yes. SleekAI reads it like any other source. For externally hosted status pages (Statuspage, Better Stack), mirror the incident posts into a WordPress CPT or feed the JSON status endpoint into the system prompt at startup, so the bot can answer 'is anything broken right now' truthfully.

SDK READMEs imported as Markdown into a WordPress CPT (or pulled into an OpenAI Files vector store) work cleanly. Fenced code keeps its language hint and indentation, so a reply about the Node SDK quotes the actual sample. Multibot lets you split per language: a Python bot, a Go bot, and a Node bot, each scoped to its own README plus the shared reference.

Yes, that is the normal setup. SleekAI is a WordPress plugin, so it runs wherever your developer portal already runs. Customers usually host the portal subdomain (developers.example.com) on the same WordPress install, with the chatbot widget attached via display conditions to /docs, /sdk, and /recipes URLs.

Multibot scopes a chatbot to URL patterns, post types, or logged-in roles, and lets you spin up several focused assistants, 'docs bot', 'recipes bot', 'release bot', each with its own scope and system prompt. A developer on /docs/webhooks gets the docs bot, a developer on /recipes gets the recipes bot, with surface-appropriate tone and source set.

SleekAI is a one-time WordPress plugin license, not a per-conversation SaaS subscription. You bring your own OpenAI, Anthropic, Google, or OpenRouter API key, so you pay the provider directly at standard API rates. There is no token markup, no resolved-conversation fee, and no monthly seat cap, which makes cost predictable even on a busy developer-portal page.

Yes, the system prompt instructs the model to end every reply with a link to the page it came from. For technical content this is non-negotiable, developers will not trust an unsourced claim about authentication or rate limits. Citations also create a feedback loop, dead or wrong links surface in logs and become tickets for the docs team.

Conversations live in your WordPress database, not on a third-party vendor's servers. Using your own OpenAI, Anthropic, Google, or OpenRouter API key with the standard API terms means request data is not used for model training by default. For SOC 2 or ISO-conscious buyers, that combination is usually easier to defend than a black-box vendor chatbot.