Editorial image showing a 3D Chatbase content-control badge with page, file, Q&A, handoff, and trust-check cues.

Website AI chatbot review

Chatbase is going Shopify-first. WooCommerce stores should plan around it.

Every chatbot purchase splits into two layers: a content layer (does it answer from your pages, files, and FAQs?) and a store layer (does it do things with your store — catalog, cart, orders?). Chatbase's content layer is strong. Its store layer shipped a native Shopify integration, five Shopify Actions, and a no-code Shopify setup blog in the last four months. None of that arrived for WooCommerce. If your job is the content layer on a WordPress site, Chatbase is the chatbot we'd reach for; if it's the store layer on Woo, you're buying half the product. Five passes, one partial, one miss on our 7-question answer test — full table below.

Updated
June 1, 2026

Chatbase is one strong layer and one half-built layer.

The content layer answers from sources you choose: pages, files, Q&A pairs, Notion, and support material. The store layer would mean catalog sync, cart actions, order lookup, customer profile updates.

We ran the same 7-question answer test we used on FastBots — a workshop-and-class business's content, the kind of mixed-content site Chatbase's content layer is built for. Five passes, one partial, one miss. Strong on content-grounded answers and unsupported-question refusal; weaker on conditional buyer questions that need a redirect (age-appropriate course, time-window refund rule). One data point on someone else's content — run the same rig on yours before relying on it.

See it in action

Chatbase dashboard + data-sources setup.

Two visual reference points before we get to the answer-quality table. Left: Chatbase's official ten-minute build-an-agent walkthrough (their channel, public). Right: the Shopify Actions docs page we cite throughout the WooCommerce caveat block. For the full vendor-captured dashboard tour, see Chatbase's own "Build Your First AI Agent" docs (9 screenshots of the actual dashboard, data sources, playground, and deploy screens). Neither is a substitute for opening a trial yourself, but together they show the surfaces this review is judging.

Walkthrough — Chatbase's official YouTube channel. Dashboard + agent + data-sources setup in roughly ten minutes.
Screenshot — Chatbase Shopify Actions docs (captured 2026-05-24). Click to expand.
Chatbase Shopify Actions documentation screenshot showing the five action entries.

Screenshot — Chatbase Shopify Actions docs (captured 2026-05-24). Click to expand.

The answer-quality check

Seven test questions. Five passes, one partial, one miss.

The same lightweight rig we used on our FastBots review — small but visible. Source material was a workshop-and-class business's public pages and policies. We asked seven questions a real customer would ask, then judged each answer against the source.

What Chatbase looks like in practice — three vendor-public screenshots from Chatbase's own docs and blog. Illustrative: these show the playground surface and what responses look like in general, not first-hand captures of our 7-question test. Verdicts in the table below are based on live observation of the agent we configured.

The test surface — from Max T, "How to Build an AI Chatbot," Chatbase blog (2025-09-24). The chat preview we tested in.
Chatbase live chatbot preview — user testing a question in the preview interface.

The test surface — from Max T, "How to Build an AI Chatbot," Chatbase blog (2025-09-24). The chat preview we tested in.

A response in practice — from Max T, "How to Build an AI Chatbot," Chatbase blog (2025-09-24). Shows response formatting.
Chatbase chatbot responding to a user question in the preview interface.

A response in practice — from Max T, "How to Build an AI Chatbot," Chatbase blog (2025-09-24). Shows response formatting.

Source-grounded answer — from same Chatbase blog (2025-09-24). Shows the content-layer behaviour we tested against.
Chatbase agent answering a domain-specific question grounded in uploaded source material.

Source-grounded answer — from same Chatbase blog (2025-09-24). Shows the content-layer behaviour we tested against.

Question type Verdict Note
Straightforward content-grounded answer Pass Asked a simple factual question whose answer lived in the imported source material. Chatbase quoted the relevant page accurately and stayed within the source.
Unsupported-question refusal Pass Asked a question whose answer was NOT in the source material. Chatbase declined to invent — said it didn't have that information.
Contact fallback Pass Asked an ambiguous question that should route to a person. Chatbase surfaced the contact path defined in setup.
Pickup / logistics nuance Pass Asked a 'can I do X in this specific situation' question. Chatbase parsed the condition and gave the right branch.
Policy facts Pass Asked about a written policy from the imported docs. Verbatim from source, no embellishment.
Age-appropriate workshop redirect Miss Asked whether a course was right for a child of a specific age. Chatbase answered the surface question and never redirected to the appropriate age band — the kind of follow-up a human host would have made.
Urgent class-change/refund caveat Partial Asked an urgent change-request question. Chatbase covered the change path but missed the specific time-window caveat that affects refund eligibility.

This is a small-N answer-quality check, not a full product test. It does not certify Chatbase against your specific store, catalog, or support workflow. Run the same rig against your own content before relying on it.

Before you trial it

Open Chatbase with these five questions in hand.

  1. What should the bot read first? Start with the pages, PDFs, FAQs, Notion notes, or support docs that answer real customer questions. Leave stale pages out.
  2. Will the first paid plan cover a normal month? Hobby currently lists 500 message credits/month. One customer chat can use more than one credit because credits are spent per AI response and model.
  3. Which store layer do you need — Shopify's five, or custom JSON? Chatbase's Shopify Actions are five named actions (get-products is the lowest-risk, update-billing-address the highest). For WooCommerce, you're writing custom API endpoints that return JSON to the agent — closer to a dev project than a setting. Decide whether you want to build that pipe before you trial.
  4. What happens when the bot is unsure? Decide whether it should collect contact details, show a safe fallback, or create a helpdesk ticket for a person to handle.
  5. What should it never promise? Write down the prices, order details, legal claims, medical claims, or service commitments that must stay with your team.

If credit math is the worry, pair this review with the message credits vs conversations guide before you compare Chatbase with a cheaper-looking plan.

Best for

Websites with help docs, files, Q&A, Notion pages, or support knowledge that should shape the chatbot's answers.

Try Chatbase free →

Not for

Stores that need a verified WooCommerce catalog, cart, or order workflow before they choose a chatbot.

Trial check

Test one content import, five real customer questions, one fallback answer, and one handoff path before depending on it.

Start the trial →

Pricing

Plan signals before you shortlist it.

Chatbase pricing is shaped by message credits, source limits, seats, and feature gates. The public pricing page reviewed here showed annual pricing, so check the monthly toggle before budgeting against it.

Free

$0/month, 50 message credits/month, 1 member, and 400 KB per AI agent. The reviewed pricing page says inactive free agents are deleted after 14 days.

Hobby

$32/month when billed annually, with 500 message credits/month, 5 enabled AI Actions per AI agent, 10 MB per AI agent, and 2 members.

Standard

$120/month when billed annually, with 4,000 message credits/month, API access, help desk, voice, telephony, auto retrain, and advanced integrations.

Pro

$400/month when billed annually, with 15,000 message credits/month, higher AI Action and data limits, advanced analytics, source suggestions, and tickets as a source.

Enterprise

Custom pricing with higher limits, custom roles and permissions, SSO, white-labeling, audit logs, priority support, CSM, and SLAs.

What you'd actually pay

Three usage profiles, three different bills.

Headline price is rarely the real bill. Credits are model-dependent, so the right plan depends on how many real customer chats you expect and how heavy each conversation is. The figures below assume Chatbase's listed annual pricing and the default per-response credit ranges we observed in the docs; recheck against your own model choice before committing.

Light use — ~50 customer chats per month

Hobby — $32/month annual ($384/year)

500 message credits cover ~50 chats if each averages ≤10 credits. Realistic for a docs site, FAQ-only support, or a low-traffic catalog. Watch for the 14-day inactive-agent deletion on the free plan if you're piloting.

Try Hobby →

Growing use — ~200 customer chats per month

Standard — $120/month annual ($1,440/year)

4,000 message credits cover ~200 chats at ~20 credits each, with API + advanced integrations + auto retrain unlocked. Budget add-ons separately.

Try Standard →

Scaled use — 1,000+ chats per month

Pro — $400/month annual ($4,800/year) plus auto-recharge

15,000 credits handle ~1,000 chats at ~15 credits each; heavier conversational flows or larger models exhaust the allowance faster. Plan on $40 per 1,000 auto-recharge credits as a top-up. Multi-agent workspaces will also need extra-agent add-ons.

See pricing →

A few credit-math rules that shape every one of those bills. Credits renew on the 1st of each month, per Chatbase's FAQ. When credits run out, the agent shows an unavailable message and usage shows up under Workspace Usage. Auto-recharge credits are listed as an add-on at $40 per 1,000 message credits on the pricing page. For the unit math behind "credits vs conversations," read message credits vs conversations before treating a 500-credit plan as 500 customer chats.

Currency is USD. Pricing captured 2026-06-03 from chatbase.co/pricing; figures may shift if Chatbase changes per-response credit math.

Setup path

How a website would actually add it.

Chatbase has a clear website and WordPress deployment path. The moment you need product, order, or cart behavior, separate the documented setup path from what still needs custom configuration or a real store trial.

Website deployment

Create and enable a Chatbase agent, add source material, then deploy the chat widget with the JavaScript embed or another supported website integration.

WordPress site

Chatbase documents a WordPress plugin path: install the plugin, paste the Agent ID into the plugin settings, and save the change.

Custom workflow

Custom actions can call external APIs and return JSON to the agent, but this is configuration work rather than a native WooCommerce workflow.

Human handoff

The escalation action creates a ticket in a connected helpdesk such as Zendesk, Salesforce, Intercom, Zoho Desk, Freshdesk, HubSpot, or Help Scout.

Content controls

The main reason to look at Chatbase.

Chatbase is strongest when answer quality depends on the sources you give the chatbot, not just how quickly you can paste a widget onto a site. If you are comparing website crawls, files, Q&A, help-center content, and product data across tools, start with which AI chatbots can answer from your website. If the decision depends on Calendly or Cal.com actions, use the appointment-booking chatbot guide to separate direct calendar actions from scheduling-link handoff.

Website content

Full-site crawl, sitemap submission, or individual links, with include and exclude path controls.

Files and snippets

PDF, TXT, DOC, and DOCX files, plus direct text snippets.

Q&A pairs

Manual Q&A entries with multiple question variations, rich formatting, and usage metrics.

Notion

Notion is documented as a source type.

Support tickets

Tickets can be used as a source when eligible helpdesk integrations are configured; the reviewed pricing page lists this on Pro.

Where to be careful

What to verify before you commit.

  • Message credits are model-dependent, so do not read 500 credits as 500 full conversations.
  • API access appears to start on Standard in the reviewed pricing and API docs.
  • Standard and Pro list auto retrain, and the current data-sources docs say those plans update the knowledge base every 24 hours. Recheck the current docs before buying mainly for that feature.
  • Custom domains appear Enterprise-only in the reviewed workspace settings docs.
  • Removing 'Powered By Chatbase' is listed as a paid add-on.

The store layer, on WooCommerce.

Chatbase has official Shopify integration and Shopify Actions docs covering product, order, cart, profile, and billing-address behavior — five named actions that our companion guide walks through from lowest-risk to highest-risk (get-products at the safe end, update-billing-address at the sharp end). There are no equivalent Chatbase docs for WooCommerce-native catalog sync, cart actions, order lookup, or customer profile updates.

Chatbase's recent product investment has been visibly Shopify-heavy. The native Shopify integration shipped in February 2026, the five Shopify Actions docs followed, then a no-code Shopify setup blog. That's a strategic direction, not an oversight. A WooCommerce store buying Chatbase today is buying the content layer — not a store layer that might ship Woo-native parity later this year. Plan around that, or choose a different tool if the store layer is the job.

The case for buying Chatbase on Woo anyway: if you have a developer who can write 3–5 Custom Actions hitting WooCommerce's REST API (products, orders, customer lookup), the store-layer gap closes for common queries. That's roughly 1–2 weeks of dev work. When it wins: your content needs are the primary job, the store-action set is small (read-only product/order lookup, not billing-address updates), and you already have engineering hours. When it doesn't: you're non-technical, or you need write actions (cart updates, address changes) where Shopify gets the official rails and Woo gets your developer's edge cases.

Either way: treat Chatbase as a general WordPress or website chatbot for the content layer; verify the store layer in your own trial before you rely on it. If you do build WooCommerce-equivalent actions, ship the safe ones first (read-only product lookup) and leave the risky ones (billing address, order changes) off until you've tested them in a sandbox — same logic Shopify buyers use when enabling Chatbase's official actions.

For the deepest Chatbase mechanics — every action, every credit rule, every integration — go to chatbase.co/docs.

Alternatives

Compare around the job, not the category.

Tool directory

Start here if you want the broader shortlist before reading individual reviews.

Read next

FastBots

Compare FastBots if you want the simpler trained-chatbot path before choosing a broader AI-agent workspace.

Read next

Tidio

Compare Tidio if live chat, support handoff, and customer-service workflow matter more than source controls.

Read next

WooCommerce guide

Use the WooCommerce guide if product questions, order issues, and store-specific caveats are the main decision.

Read next

FAQ

Chatbase FAQ

What does Chatbase do well?

We found Chatbase strongest at controlling the source material an AI agent reads before answering. The pricing page shows a free plan with 50 message credits/month and 400 KB per AI agent, then larger source quotas on paid plans. That setup helps teams that care more about which pages, files, and Q&A pairs the bot trusts than about a live support inbox. For a side-by-side with a simpler trained bot, read the FastBots vs Chatbase comparison.

Reviewed · Sourced from Chatbase pricing

Where does Chatbase struggle?

We saw two friction points. First, the free plan deletes inactive agents quickly: the pricing page says AI agents are deleted after 14 days of inactivity on the free plan, so a paused project can vanish. Second, removing platform branding and adding extra agents are paid add-ons, so the headline plan price is not always the final bill. Recheck add-on costs and storage limits in your Chatbase workspace before treating the entry tier as enough.

Reviewed · Sourced from Chatbase pricing , Chatbase pricing add-ons

Is Chatbase a good fit for a WordPress site?

For a WordPress site that mostly needs answers from pages, files, and FAQs, Chatbase is a sensible first trial. The official WordPress setup doc walks through finding the Chatbase plugin from inside WP Admin and pasting the Agent ID into the plugin settings, which is a normal WP install path. If you need WooCommerce order or cart actions, treat Chatbase as a content-answer layer rather than a store automation layer and confirm with your own test.

Reviewed · Sourced from Chatbase WordPress setup

How does Chatbase pricing work?

Chatbase pricing is shaped by message credits, AI agent storage, members, and add-ons. The pricing page lists Hobby with 500 message credits/month, Standard with 4,000 message credits/month, and Pro with 15,000 message credits/month, plus larger source storage at each tier. On top of that, auto-recharge credits, extra agents, and branding removal are listed as add-ons. Estimate real monthly replies on your chosen model first, then add the add-ons you would actually buy.

Reviewed · Sourced from Chatbase pricing tiers , Chatbase pricing add-ons

What are the best alternatives to Chatbase?

We compare Chatbase against three different jobs. For a simpler trained website bot, FastBots is the cleaner side-by-side: its WordPress page says you install the official plugin, paste a Source URL and Embed ID, and go live. For live chat and tickets, Tidio fits better, and its pricing page lists billable conversations plus a separate Lyro AI Agent quota. For a deeper feature view, read the FastBots vs Chatbase comparison or the WooCommerce chatbot guide.

Reviewed · Sourced from FastBots WordPress integration , Tidio pricing

Decision recap

Chatbase fits if...

  • Buy the content layer on Hobby — if the bot's job is answering from pages, files, Q&A, Notion, and help content — and 500 credits covers your month.
  • Buy the content layer on Standard — if you need API access, AI Actions, auto retrain, and helpdesk escalation on top — and 4,000 credits covers your month.
  • Skip Chatbase for the FastBots comparison — if the content layer is your whole job and a simpler trained-site bot might cover it for less.
  • Skip Chatbase for the Tidio comparison — if the team's real job is live chat, tickets, and human handoff — not a content-grounded bot.
  • Wait on the store layer — if a WooCommerce store needs proven catalog, cart, or order behaviour today — Chatbase's store layer ships for Shopify, not Woo.
Page changelog
  • 2026-06-05 — Added the two-layer framework, three TCO scenarios, and a sharper take on what Chatbase's Shopify-heavy direction means for Woo buyers.
  • 2026-06-01 — Original publish.