Documentation-based evaluation, checked September 24, 2026. We have not completed a live support-agent, integration, or handoff test. The fictional cases below describe a proposed evaluation, not measured Chatbase performance.
| Plan | Monthly price | Message credits per month |
|---|---|---|
| Free | $0 | 50 |
| Hobby | $40 | 700 |
| Standard | $150 | 4,000 |
| Pro | $500 | 15,000 |
| Enterprise | Custom quote | Confirm in quote |
My short verdict
Choose by the support journey you can verify. Start with routine questions, then test unknowns, outdated policies, and explicit requests for a person. A chatbot that gives attractive answers but loses escalations is not ready to own that journey.
It makes sense when
- A business with a current help center and named content owner.
- A support team that can review transcripts and receive escalations.
- A buyer willing to validate required integrations on the eligible plan.
I would choose something else when
- Your policies conflict and nobody can approve the correct answer.
- You need verified autonomous actions without integration testing.
- You expect a vendor resolution percentage to predict your own results.
01
What Chatbase offers
Chatbase describes a support platform that uses supplied knowledge, connected actions, and human escalation. Its product overview includes monitoring conversations and improving instructions or sources. These capabilities identify what to evaluate; they do not prove accuracy on your policies.
For a small store, separate questions about return rules from requests to refund an order. The former can be answered from approved content. The latter needs an authorized operational process. An assistant saying an action is complete is not proof that a system performed it.
02
Read the plan against your required workflow
Standard's checked card includes Helpdesk, voice, telephony, API access, and auto-resync. Check the exact channel and integration you need before selecting a tier. The free plan's agents are deleted after 14 days of inactivity; a trial is not a permanent unattended support deployment.
The table uses monthly billing. Compare annual commitments separately and confirm current checkout and renewal terms. Choose based on needed features, then measure usage with representative questions and follow-ups. Do not translate credits directly into a promise of resolved tickets.
03
Optional costs can change the decision
The pricing page lists extra agents at $25 each per month, branding removal at $99 per month, and recharge at $40 per 1,000 message credits. Confirm eligibility and configuration before selecting add-ons; none is assumed necessary for every buyer.
A hypothetical monthly configuration of Standard plus one extra agent and branding removal totals $274 before usage extras and applicable tax: $150 + $25 + $99. This is arithmetic from the listed options, not our bill or a recommendation to buy them.
Also count staff time for reviewing failures, updating policies, and maintaining integrations. A lower subscription can still require more operational work. Compare accepted answers and completed handoffs, not merely the amount of text generated.
04
Use a fictional help center before real customers
Our proposed shared support evaluation uses a fictional desk-organizer store. Its current policy allows unused-item return requests within 30 calendar days after delivery and free shipping from a $60 merchandise subtotal before tax. An explicitly archived policy says 14 days and $75. The agent must distinguish the versions.
The test includes 30 questions: ordinary policy answers, threshold calculations, missing product facts, unavailable order lookup, cancellation requests, after-hours escalation, and attempts to override the rules. No real customer or payment details are needed.
For example, ask whether an unused item delivered 20 days ago can be returned. A useful answer identifies the window and human approval requirement without claiming a refund was issued. Ask about an unknown order and check that the agent does not invent shipment information.
- Compare each response with the approved answer key and retain the full transcript.
- Check whether the agent clarifies ambiguous requests instead of guessing.
- Verify a handoff arrives in the designated test inbox with the conversation context.
- Ask for a person explicitly and check that the conversation does not loop through more bot answers.
- Record plan, model, channel, source settings, credits, and review time.
05
Test what happens when a policy changes
Change the fictional free-shipping threshold from $60 to $65, update the source through the documented workflow, and repeat the affected questions. Keep the original responses. Record when the new policy becomes usable and whether the agent still repeats the archived value.
For production, give someone ownership of the source collection and an escalation queue. A source refresh is only useful if the correct information reaches answers. Review unresolved requests and repeated misunderstandings before expanding to more channels.
06
How to compare Chatbase with another support tool
Keep the same corpus, questions, required channel, and receiving inbox. Compare eligible plans first; a plan without the required integration is not a like-for-like price comparison. Separate unsupported features from broken implementations.
Measure factual accuracy, appropriate refusal of unsupported actions, successful handoffs, and correction effort. A closed conversation can mean the customer left; it does not automatically mean the issue was solved. We have not completed this paired test, so this review does not claim Chatbase beats another platform.
Ready to test it with a real job?
Check which plan supports your required workflow, then evaluate one representative task. Record output quality, correction time, and actual usage before expanding.
Check Chatbase plansSponsored affiliate link.
Questions people ask before paying
Is Chatbase free?
A free plan is listed with 50 monthly message credits. Its agents are deleted after 14 days of inactivity. Check whether the features required for your evaluation are included.
Do message credits equal customer conversations?
Do not assume a one-to-one relationship. A conversation can involve follow-ups and different operations. Verify current consumption rules and measure your representative workload.
Can Chatbase hand conversations to a person?
Human escalation is part of its documented product offering. Verify the eligible plan, destination, transcript continuity, and actual receipt in a test inbox; we have not completed that test.
Are these annual prices?
The comparison table shows displayed monthly prices. Review annual billing and renewal commitments separately before subscribing.
Has this review tested Chatbase with real customers?
No. This is a documentation-based evaluation with a proposed fictional support benchmark. No accuracy score, saved-time figure, or live-customer result is claimed.
Sources and further reading
Pricing and feature details can change. Consult the linked product documentation for current details and check the evidence note for the scope of this evaluation.