01
Sources on every answer
Answers are grounded in your published articles and cite them inline. Readers verify, click through, and keep reading — the AI sends traffic to your archive, not away from it.
The AI archivist for WordPress publishers Est. 2026 · No. 001
A WordPress plugin for media websites
Years of articles, buried by the scroll. Archivist AI turns your WordPress archive into a conversational search engine — readers ask, your articles answer, with sources cited.
What it does
Archivist AI is not a generic assistant. It answers only from your own published content — and always shows its work.
01
Answers are grounded in your published articles and cite them inline. Readers verify, click through, and keep reading — the AI sends traffic to your archive, not away from it.
02
"Last month", "in 2024", "since January" — questions with dates are parsed into real date filters before searching, so news archives answer like news archives should.
03
Visitors chat anonymously — no login, no signup wall, no email capture. Abuse protection works per-IP, never by forcing your readers to register.
04
A cost estimate before indexing. A dashboard for every cent. A daily spending cap with graceful fallback. You pay OpenAI directly at list price — we never mark up your AI costs.
05
Install, paste one OpenAI API key, click Index. Embeddings live in your own WordPress database — no vector database accounts, no extra services, works on any host.
06
Answers appear word by word, with a floating bubble on every page — or embedded inline anywhere via shortcode and Gutenberg block.
Setup
1
Install from WordPress.org and add your OpenAI API key. That is the only credential in the entire plugin.
2
See the estimated cost before anything is spent — a 10,000-article archive indexes for a dollar or two. New posts index automatically.
3
A chat bubble appears on your site (or embed it inline with a block). Readers start asking. You watch the dashboard.
Radical cost honesty
Most AI plugins hide the meter. Archivist AI shows you the estimate before indexing, logs every API call, and stops at the daily cap you set. You pay OpenAI directly on your own key — we take nothing on top.
Approximate, at OpenAI list prices. The plugin estimates against your real archive.
| What | Volume | Cost |
|---|---|---|
| Indexing (one-time) | 10,000 articles | $0.30 – $2 |
| Reader chat | per 1,000 turns | $1 – $3 |
| Small magazine, monthly | ~1,000 chats | ~$1 – $3 |
| Mid-size portal, monthly | ~10,000 chats | ~$10 – $30 |
| Archivist AI markup | always | $0 |
Pricing
Free
$0
forever, on WordPress.org
Pro
$99/yr
single site · 30-day money-back
Questions
No. Readers chat anonymously — no login, no signup wall, no email capture. Rate limiting works per IP address, never by forcing visitors to authenticate. (You can optionally restrict the chat to your subscribers as a premium perk — that is your choice, not ours.)
You pay OpenAI directly on your own API key at list price, with no markup. Indexing a 10,000-article archive costs roughly $0.30–$2, one time. Ongoing chat costs are about $1–$3 per 1,000 chat turns. The plugin estimates costs before indexing, tracks every cent, and lets you set a hard daily cap.
No. Embeddings are stored in a custom table in your own WordPress MySQL database. Nothing extra to sign up for, configure, or pay for — it works on any standard WordPress host.
No. Only published, public, non-password-protected posts are ever indexed, and post visibility is re-checked at answer time before anything reaches a reader.
OpenAI, with your own API key — one key covers both indexing and answers. Support for additional providers is on the roadmap.