Chatbase Pricing Plans: Hobby, Standard or Pro for a Small Site

For a small website, start with Chatbase Hobby at $32 a month billed yearly and 700 monthly message credits if you need an AI agent and basic analytics. The tiers once called Standard and Pro are now shown as Basic at $120 and Business at $400, with 4,000 and 15,000 credits. Choose by observed use and handover needs.

A message credit is not necessarily one completed customer conversation. The selected model can consume more than one credit per response, and a visitor may ask several questions. Chatbase currently lists auto-recharge at $40 per 1,000 extra credits. Build a forecast from responses per conversation, model cost and the number of conversations you expect, then verify it in the workspace Usage tab.

Translate the old plan names before comparing prices

Some Chatbase pages and older comparisons still say Standard and Pro: Chatbase's own 7 October 2026 changelog entry on automated tests gives 500 test credits on Standard and 1,000 on Pro, the same allowances the pricing page now gives Basic and Business. The live pricing page shows Hobby, Basic, Business and Enterprise. Its Basic tier carries the 4,000-credit allowance associated with Standard in older materials; its Business tier carries the 15,000-credit allowance associated with Pro. Keep the assigned title in mind when searching, but use the plan names and checkout terms presented to your account at purchase. Do not mix a current annual price with an old plan's feature list.

The rates above are monthly equivalents when billed yearly. That means a full-year commitment, not a one-month trial at the annual rate. With the page's toggle set to monthly billing, the same tiers show $40, $150 and $500 a month; the yearly view labels its savings as $96, $360 and $1,200 a year, which matches the 20% discount the toggle advertises. Annualised base totals are $384 for Hobby, $1,440 for Basic and $4,800 for Business, before add-ons or extra credits. Paying monthly for a year would cost $480, $1,800 and $6,000. Confirm the figure in the order summary before you pay.

Chatbase lists a seven-day trial on these paid tiers and a separate test-credit allowance: 200 tests on Hobby, 500 on Basic and 1,000 on Business in the live comparison. Test credits are not the same thing as visitor message credits. Use them to check an agent before publishing, then inspect the live credit ledger after launch. Chatbase's own automated-test notes say paid plans receive test credits; do not assume every internal test uses the visitor allowance or that tests are free without limit.

The Complete Excel Course with AI Integration

From formulas and pivot tables to Power Query, DAX, VBA and AI with Copilot, ChatGPT and Claude — hands-on from the first lesson.

By Sagnik Bhattacharya and Virginia Thorn, Primetime Emmy Award winner

Buy now

Convert visits into response credits

The shortest useful forecast has four inputs: website visitors per month, the percentage who open the chat, average AI responses per chat and credits per AI response for the chosen model. If 2,000 visitors arrive, 8% start a conversation, each conversation uses three AI responses and each response costs two credits, the monthly use is 2,000 × 0.08 × 3 × 2, or 960 credits. Those numbers are illustrative assumptions, not a Chatbase benchmark. Hobby's 700 included credits would be short by 260.

Actions change the arithmetic. Chatbase's pricing FAQ says each request to the AI model costs credits according to the chosen model, and an agent with Actions enabled may make several requests to produce one reply. Its own example is a calendar booking: one request decides to run the calendar Action and a second writes the answer, so the reply costs twice the model's rate. If the print shop's bot checks order status or books calls, count those replies separately in the forecast.

There is a second way to express that forecast that helps an owner compare plans. At three responses costing two credits each, one conversation uses six credits. Hobby's 700 included credits would cover roughly 116 such conversations, leaving a small unused remainder. Basic's 4,000 would cover roughly 666, and Business's 15,000 would cover 2,500. These are mathematical capacities under the assumed conversation pattern, not traffic predictions or guaranteed resolutions. If a visitor asks one question, more chats fit; if a bot repeatedly asks for clarification, fewer fit.

For an illustrative margin, the print shop could reserve 20% of its allowance for an unexpected campaign. Hobby then has 560 planned credits rather than 700, while Basic has 3,200 rather than 4,000. That margin is an internal planning choice, not a Chatbase rule. It can prevent a small burst of visitor questions from exhausting the bot on the busiest day. If the shop wants certainty beyond that margin, it should test whether auto-recharge or a visible fallback form fits its customer promise.

At Chatbase's published $40 per 1,000 auto-recharge credits, a 1,000-credit add-on could cover that illustrative shortfall if the account offers that purchase path. The annual-billing Hobby base of $32 plus one $40 recharge purchase would mean $72 paid in the month the top-up occurs. The unused 740 top-up credits are not the same as that month's included credits: Chatbase says auto-recharge credits do not expire, while subscription credits renew at the start of each calendar month, whatever day you subscribed, and unused ones do not carry forward. Do not call $72 the permanent monthly cost unless usage and purchase timing actually repeat.

Now try a more active print shop site. It receives 420 chat conversations a month, averaging four AI responses at one credit each. That is 1,680 credits. Hobby would need more than its 700 included credits; a single 1,000-credit purchase could cover the difference for the month, costing a $72 cash outlay at the annual-billing base plus recharge. Basic at $120 includes 4,000 credits and a Helpdesk inbox. The price difference is $48 in that illustrative month, but Basic may be the better fit if staff need to take over conversations within Chatbase. Volume alone does not decide it.

If the agent uses a model that charges two credits per response, the same 420 × 4 conversations consume 3,360 credits. Hobby would require three 1,000-credit top-ups beyond its included 700 to cover the full use, or $120 in recharge purchases and the $32 base in that month. Basic's 4,000 included credits would cover the illustrative volume for its $120 annual-billing monthly equivalent. This demonstrates why the model choice can reverse the cheap-plan answer. Chatbase's 27 October 2026 changelog entry, for example, prices one newly added model at 2 credits per response; check the model setting for your actual agent.

Consider the annual view as well. If the print shop has six ordinary months at 1,680 credits and six busy months at 3,360 credits, its total annual use is 30,240 credits. Hobby includes 8,400 subscription credits across twelve months, but unused credits from one month do not cover a busy month later. Basic includes 48,000 across the year, yet that headline total also hides monthly boundaries. Model each month separately, then add actual top-up purchases. A single annual average would make the seasonal peaks look smaller than they are.

Do a break-even calculation only after you know the feature need. Basic costs $88 more than Hobby as an annual-billing monthly equivalent. At the stated $40 per 1,000 top-up, three top-ups bought in a month cost $120, more than that $88 difference. But a top-up can leave non-expiring credits for later, while a Basic subscription's monthly allowance resets. There is therefore no permanent break-even at one fixed chat count. Compare twelve monthly forecasts, carry forward the unused purchased balance and include the value of the Helpdesk only if the team will operate it.

When Hobby is enough for a site

Hobby suits a controlled, low-volume answer tool with two team members, basic analytics, one agent and a modest approved knowledge base. Chatbase's current comparison lists 700 monthly message credits, 200 monthly test credits, one agent, five AI actions and a 40 MB training-content size. An action is a step the agent can perform beyond composing text; whether a particular action is available and safe for your website needs testing. Hobby does not show the built-in Helpdesk inbox. If a visitor must speak to a person, design a separate route that your team can actually monitor.

For example, a yoga studio may want answers to public class times, membership options and what to bring. The owner sets a rule that the bot may quote only published pages, while booking changes go to a staffed form. During an illustrative quiet month it receives 90 conversations with two one-credit AI responses each, or 180 credits. Hobby has ample headroom. The studio should still test a visitor asking for a refund or a medical recommendation; those questions should go to a person rather than being answered from vague content.

A different low-volume example is a wedding planner's portfolio site. It might see only 30 chats in a month, but each enquiry could involve multiple questions and a handover. If the business needs a staffed inbox with chat history and assignment inside Chatbase, Basic's feature may matter despite the low credit count. If the website already sends enquiries to a reliable form and CRM, Hobby can answer common questions and direct people there. Do not pay for Helpdesk merely because the bot can collect names; buy it when the team will use the workflow.

Basic earns its place through human follow-up

The current Basic plan adds the Chatbase Helpdesk inbox, advanced integrations, voice, API access, automatic source refresh and other features listed on its live page, with three members and 4,000 monthly credits. Those are plan features, not a promise that every website will use them. The operational trigger is a question that AI should not finish alone. Can a person see the visitor's original question, take ownership, respond within the business's promised time and close the case? Pilot that path before paying for a larger credit bucket.

An IT support firm could ask the bot to answer opening hours and public service scope while handing technical incidents to a person. Suppose 300 visitors ask three two-credit responses each: 1,800 credits. Basic's 4,000 included credits cover the illustrative month, but the Helpdesk should be tested with the actual duty rota. If a visitor asks whether a damaged server is under contract, the bot must not promise coverage based on generic service copy. It should collect a case reference and pass it to staff. The handover tutorial gives the operating rule in more detail.

Set the handover policy in writing: “Escalate refunds, complaints, disputed prices, account-specific information and any request that the source material does not answer.” A print shop may add “artwork deadlines and colour-match guarantees”. One visitor may ask three routine questions and then a custom quote question; the agent's earlier credits have already been used even if the final request goes to a person. Measure the cost per useful handled conversation, not just cost per chat opened.

Business needs a real reason beyond more credits

Chatbase's Business tier has 15,000 monthly message credits, five members, advanced analytics, source suggestions and tickets as a source, according to the live comparison. It makes sense when both volume and oversight justify it. A small site with 2,500 short conversations using two one-credit replies each would consume 5,000 credits. That is over Basic's 4,000. One $40 top-up of 1,000 credits may be much cheaper than moving from Basic's $120 base to Business's $400 base. If advanced analytics or the larger team capacity matters, the comparison changes.

Suppose a print shop has 4,000 conversations with three two-credit replies each, or 24,000 credits. Even Business's 15,000 credits would be short by 9,000, potentially nine $40 top-ups at the stated add-on size, or $360 in purchases during that month. Add the $400 annual-billing monthly base and the illustrative cash cost becomes $760. That amount is not a forecast for all print shops: the conversation count, length, model and purchase timing are assumed. It shows that a high-volume site can outgrow a plan's headline allowance even on the top self-service tier.

The owner should ask whether every conversation needs the expensive model. Many basic questions can be answered from a clear FAQ, a site navigation link or a simple one-credit response if the available model handles them accurately. A model change should be evaluated against answer quality as well as credit cost. If a cheaper choice produces more repeated questions or wrong answers, the apparent saving disappears. The performance measurement tutorial explains how to count resolved, escalated and incorrect conversations.

Account for the add-ons that change the bill

Chatbase's current price page lists auto-recharge at $40 per 1,000 message credits, an extra agent at $25 per month, extra storage at $25 per 10 MB per month and removal of “Powered By Chatbase” branding at $99 per month. Each is a separate choice. The pricing FAQ adds that the 40 MB allowance per agent is the size of extracted text, not of your files, and that the storage add-on can add up to 40 MB more in 10 MB steps; beyond that, Chatbase points you to Enterprise. A business that wants two distinct agents, more source capacity and no branding could pay substantially above the base tier even while message use stays low. Write each add-on into the budget only if the site truly needs it.

A filled-in illustrative budget for a small IT support site might be: Basic base $120 monthly equivalent on annual billing; one extra 1,000-credit recharge purchase in a peak month, $40; no extra agent; no branding removal; staff review time, two hours at the firm's own internal cost. The cash outlay for vendor items in that peak month is $160. The internal review cost cannot be priced without the firm's own rate, but it should not vanish from the decision. An annual plan is still a commitment even when this illustration calls the base a monthly equivalent.

Another business may turn on auto-recharge for continuity but forget to set an operational approval rule. If website traffic spikes because a bot enters a loop with visitors, the balance can be consumed by repeated replies. Put a threshold and a named owner in the billing settings; check usage by agent weekly at first. Chatbase's workspace Usage tab shows credits used in the period you choose, broken down by agent (including deleted agents), with the day-by-day trend. A credit limit can protect the budget but can also stop the bot, so the site should show a reliable contact route when AI is unavailable.

Test five visitor questions before committing

Build a test set around real site copy: hours, service area without implying an unavailable location, a price question, a custom request, a complaint and a question with no published answer. For each, mark the source page, the response you would accept and whether a person should take over. An illustrative wrong answer is a bot saying “All custom print orders arrive tomorrow” because an old promotional page mentions next-day delivery. Remove or update that source and retest. Do not upgrade to Business in the hope that a larger credit allowance will correct obsolete content.

Count the credits used by the test set and by a week of live pilot traffic. If 100 pilot chats consume 260 credits, the observed average is 2.6 credits per chat. Projecting to 300 similar chats would suggest about 780 credits, just above Hobby's 700. Add a margin for longer questions, then choose between a controlled top-up and Basic. The projection is valid only if the future traffic mix resembles the pilot. Record unusual campaigns or seasonal peaks separately.

Finally, verify the handover and lead record. A visitor who asks for a quote should receive a clear route to leave details, but the bot should not invent a delivery date or promise a price. Try a fake enquiry with a missing phone number, then inspect where it lands and who follows up. If your existing form handles it better, keep the bot focused on public questions. If Chatbase's Helpdesk improves the path and staff actually use it, that is a stronger reason for Basic than a vague desire for a “professional” plan.

The plan decision can now be written as one sentence: “Use Hobby while our observed credits fit 700 and the existing form handles handover”; “Use Basic because a staffed inbox and 4,000 credits meet our tested demand”; or “Use Business because higher volume and its oversight features are used”. Recheck the live plan names and checkout before payment, since older Chatbase material still uses Standard and Pro. For the separate choice of Chatbase versus ChatGPT on a public site, assess whether you need a visitor-facing widget at all.

Further reads

• How Much Does a Website AI Chatbot Cost, and Will It Pay Off? — Compare chatbot cost with useful outcomes.
• What to Ask an AI Chatbot Vendor Before You Sign Up — Check the contract before purchase.
• How to Train an AI Chatbot on Your FAQs, Policies, and Prices — Prepare accurate source material.
• How to Test a Customer Chatbot Before It Goes Live — Run a launch test with real questions.
• 9 AI Chatbot Mistakes That Lose Small Businesses Customers — Spot failure patterns in visitor chats.
• Roofing Chatbot Platforms With AI Compared: Qualifying Storm Enquiries — No roofing chatbot wins on its vendor page alone: the ten questions a storm bot must ask, nine platforms compared, and a ten-enquiry test to run first.
• AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.

Sources: Chatbase pricing page, plan comparison and pricing FAQ (chatbase.co/pricing); Chatbase changelog entries of 7 and 27 October 2026 on automated tests and model credit costs (chatbase.co/changelog); Chatbase Usage documentation (chatbase.co/docs/user-guides/workspace/usage). The live pricing page checked 5 November 2026 uses Basic and Business where Chatbase's own October changelog still says Standard and Pro.