
CuServly
CuServly is a no-code customer service chatbot that learns your business from your own web pages and documents, then answers visitor questions in 95 languages, with unlimited messages and an optional handover to a human agent.
What is CuServly?
CuServly is a customer service chatbot published by PiconAI, a one-person Croatian company. Its premise is narrow and deliberate: rather than drawing on general knowledge, the bot answers only from material you supply. You point it at your website's URLs, upload documents in PDF, DOC, TXT, PPT, XLS or CSV, and it builds a knowledge base out of them.
What sets it apart from most chatbot builders is that this knowledge base is not sealed. It appears in the portal as question-and-answer units or as chapters, depending on the source, and you can add, edit or delete any of them. The publisher makes that visibility the centre of its pitch, and the claim holds up: you can see precisely what your bot will tell customers, and correct it before they hear it.
Setup targets people who do not code. Three steps, appearance then data sources then embedding, and the publisher claims two minutes. The bot goes onto your site as a widget, an iFrame or a full-page window, each bot getting its own URL, and WordPress users paste a JavaScript snippet into any custom-code plugin. Every plan carries unlimited messages on unlimited websites, so the bill tracks the number of bots and the volume of source material rather than your traffic.
The multilingual claim is the headline: 95 languages understood and answered from one knowledge base. Beside it sit modest analytics, the five most-asked questions on the dashboard and per-question frequency in the knowledge base view, plus a Human Assistance option letting a live agent take over, reserved for the top plan. Customisation covers colours, avatar, widget colour, icon, logo, bot messages, quick prompts and the line shown when the bot does not know.
Technically it is a hosted web application. Data sits encrypted on Google Cloud, and when a question does not match a stored entry, relevant chunks go to a third-party AI model to generate the reply. There is no API and no mobile app.
One fact overrides the rest: on 11 August 2026 the homepage carried a hard-coded outage notice and the customer portal returned HTTP 503. That notice was absent from the February 2026 web archive, so the interruption is recent.
What it does
- Answer customer questions automatically, in 95 languages
- Train a chatbot on your own website URLs and documents
- Inspect, edit and delete everything the bot knows, entry by entry
- Hand a conversation over to a human agent when the bot falls short
- Embed the bot as a widget, an iFrame or a full-page window
- See the five most-asked questions and how often each one comes up
- Customise colours, avatar, logo, quick prompts and the fallback message
When to use CuServly / When not to
A quick filter to help you decide if CuServly is the right fit.
When to use CuServly
- Small businesses and solo founders who field the same website questions over and over and have no developer to build a bot
- Support teams that want to see and correct exactly what the bot knows instead of trusting a black box
- Multilingual sites, where a single knowledge base answers in 95 languages with no separate setup per market
- Hotels, travel agencies, online shops, clinics, banks and HR teams, the six use cases the publisher puts forward
- Sites with heavy traffic but a modest budget, since every plan includes unlimited messages on unlimited websites
When not to use CuServly
- Teams that need an API: none is published, and there is no technical documentation at all
- Anyone expecting a native mobile app, as there is neither an iOS nor an Android application
- Large documentation sets: even the top plan stops at 500 web pages, 10 files and 5 MB per PDF
- Organisations that require a contractual data residency commitment, because no hosting country or region is guaranteed
- Buyers who need the service running today: on 11 August 2026 the site showed an outage notice and the customer portal returned HTTP 503
How to use CuServly
A typical end-to-end flow, from setup to results.
- Create an account on the CuServly portal, or sign in with a Google account
- Start the seven-day free trial, which asks for no credit card
- Set the bot's identity: name, colours and logo
- Add your data sources: website URLs to crawl, and files in PDF, DOC, TXT, PPT, XLS or CSV
- Let the bot train on that material, remembering that the more you supply, the better it answers
- Open the Knowledgebase view and check what it actually learned
- Add, edit or delete knowledge units until the answers are right
- If crawling produced poor results, switch on Advanced parsing mode or replace the URL with a hand-built data file
- Pick an integration format: widget, iFrame or full-page window
- Paste the JavaScript snippet from the portal's Integration - Widget section into your site, or into a custom-code plugin on WordPress
Pros & Cons
Pros
- You can read and rewrite everything the bot knows, which is rare transparency in this category
- Unlimited messages on every plan, including the cheapest one
- 95 languages with no extra configuration and no per-market setup
- Seven-day trial with no credit card requested
- A complete, public legal pack: terms, privacy, cookies, refunds and a genuine data processing agreement
- Known questions are answered straight from your knowledge base, without a round trip to the third-party model
- Prices published openly, monthly and yearly, with 17% off annual billing
Cons
- The service was unavailable when we checked: an outage notice on the homepage and HTTP 503 on the portal
- No API and no technical documentation, so no programmatic integration is possible
- No mobile application, on either platform
- Low volume caps, with the entry plan allowing 5 web pages and a single data file
- No conversation history at all on Basic, and human handover only from EUR 129 per month
- No hosting country or region is committed to, and the subprocessor list travels only on request
- Legal documents untouched since January 2024 and visibly templated, the privacy policy citing Greek law for a Croatian company
Pricing & Plans
There is no permanent free plan. CuServly offers a seven-day free trial that requires no credit card, after which the lowest paid entry point is the Basic plan at EUR 29.00 per month, or EUR 290.00 per year. Prices are displayed exclusive of applicable taxes and fees, and payment is processed through Stripe.
- 1 chatbot
- 5 web pages
- 1 data file
- 100 KB per text file (TXT
- DOC
- XLS
- CSV)
- 500 KB per document (PDF
- 2 chatbots
- 50 web pages
- 5 data files
- 500 KB per text file
- 2 MB per document
- unlimited messages on unlimited websites
- 30 days of conversation history
- no Human Assistance
- 10 chatbots
- 500 web pages
- 10 data files
- 2 MB per text file
- 5 MB per document
- unlimited messages on unlimited websites
- unlimited conversation history
- Human Assistance included
Data, GDPR & hosting
A consolidated view of how CuServly handles your data.
GDPR overview
GDPR compliance is claimed explicitly: the privacy policy invokes Regulation (EU) 2016/679, and a data processing agreement is published openly rather than offered on request. Roles are set out properly, controller for account holders and processor for customers of the application. Rights under Articles 15 to 22 are exercised at support@cuservly.com, with a one-month deadline extendable by two. Transfers outside the EEA rely on the European Commission's standard contractual clauses, and Schrems II is acknowledged. Subprocessors are covered by a list shared on request, with 30 days' notice before any change and a right to object. No Article 27 representative is named, which is consistent with an EU-based Croatian publisher. One caveat: the policy cites Greek statutes and the Greek supervisory authority, a template artefact its author never corrected.
Who owns the data?
PiconAI, a sole trader based at Oprić 80a, 51415 Lovran, Croatia, operates CuServly. Its role depends on who you are: it is the controller of your own account details, but only a processor for the content and conversations you run through the application, where you remain the controller. The data processing agreement binds it not to retain, use or disclose that data for any purpose beyond the agreed one, and forbids it from buying or selling it. Content you upload stays yours: you grant nothing more than a non-exclusive, royalty-free licence limited to operating and maintaining the service. A single address, support@cuservly.com, handles every request.
Reuse rights
For its own account data, PiconAI relies on consent, performance of the contract, legal obligation and legitimate interest, and collects name, e-mail, Google sign-in details, billing address and VAT number, plus browser, operating system and IP data. For the chatbot itself the customer decides what goes in, and can add, edit or delete any knowledge base entry at will; nothing in the terms requires you to ask permission before reusing your own content elsewhere. Two flows leave the publisher's systems: Stripe processes payment, and when a visitor's question does not match an entry already stored, relevant chunks are sent to a third-party AI model to compose the answer, with the FAQ pointing to OpenAI's privacy notice for that step. Everything else sits encrypted on Google Cloud, and the publisher states it neither shares nor sells it. Google Analytics is opt-in and switched off by default.
Data retention & training
Hosting summary
Data is stored encrypted on Google Cloud, and that is as precise as the site ever gets: no hosting country and no region is committed to anywhere in the documentation. The privacy policy states that transferring data outside the EU or EEA is not the intention, while conceding that the location or multinational nature of some providers may cause it, in which case the European Commission's standard contractual clauses apply, with Schrems II explicitly acknowledged. The data processing agreement goes further, barring transfers out of the EEA unless an adequacy decision, binding corporate rules or standard contractual clauses are in place, and requiring the publisher to assess the destination country beforehand. Named providers include Stripe for payment, Google Cloud for storage, and Google and Facebook tools for marketing and analytics. For reference, the marketing site itself resolves to 178.218.165.202, hosted in Croatia by Plus Hosting Grupa, though that concerns the website rather than customer data. Buyers with a residency requirement should ask for a written commitment, because the published documents contain none.
Where CuServly works
Country-level availability.
Not available in
Things to keep in mind
Risks and trade-offs to weigh before adopting CuServly.
- Check that the service is running before you pay: on 11 August 2026 the site announced an outage and the portal returned HTTP 503
- Do not read the FAQ's "Certainly" as a free plan, because there is none, only a seven-day trial
- Anything you load into the knowledge base may be sent in fragments to a third-party AI model to compose answers, which is worth weighing before adding confidential material
- A bot that answers only from your own content will confidently repeat whatever is wrong in it, outdated prices and policies included
- No data residency commitment and no published subprocessor list make this hard to slot into a strict procurement review
- There is no documented way to exclude your data from model training, and no security certification of any kind
- The publisher is a single-person company, so continuity of service and support rests on one individual
Setup & Integrations
Technical difficulty
Low. Building and training the bot requires no code at all: three steps, and the publisher claims two minutes. The only technical moment is putting the bot on your site, by pasting a JavaScript snippet or using an iFrame; on WordPress a custom-code plugin does the job. Each bot also has its own URL, so embedding can be skipped entirely. Expect to spend your time on content quality rather than installation: if crawling your site yields poor answers, you switch on Advanced parsing or build the knowledge file by hand. With no API, there is nothing to integrate programmatically.
Deployment
Integrations
Supported languages
Behind CuServly
Resources
All the official URLs gathered for verification and reference.
Frequently asked questions
What does CuServly train on?
How many languages does it handle?
Is there a free plan?
Can I correct what the bot says?
How does the bot get onto my site?
Where is my data stored?
Does my data leave the publisher's systems?
Is there an API or a mobile app?
Is a data processing agreement available?
Is the service actually running?
Should you pick CuServly?
CuServly does one thing and states it plainly: it turns your website and your documents into a multilingual FAQ bot, and lets you read and rewrite everything it learned. That last point is its real distinction. Most chatbot builders hide the knowledge base, whereas CuServly puts it in the portal as editable units, which is what separates a bot you can put in front of customers from one you merely hope is right.
The rest of the package is coherent for a small business. Unlimited messages on every tier, 95 languages without per-market configuration, a seven-day trial with no card, published prices, and a legal pack covering terms, privacy, cookies, refunds and a real data processing agreement that many larger vendors never put online at all.
The limits are just as clear. No API, no mobile app, and volume caps that stay modest even on the EUR 129 tier. Nothing is promised about where data lives, the subprocessor list travels only on request, and no security certification is claimed. The legal documents have not been touched since January 2024 and carry visible template residue: a Croatian publisher's privacy policy citing Greek statutes, a data processing agreement naming an unrelated company.
Then comes the fact that outweighs any feature comparison. On 11 August 2026 the homepage carried a hard-coded outage notice and the customer portal answered HTTP 503. The notice was not there in the February 2026 archive, so this is a recent interruption rather than a long abandonment, and the publisher says it will be back. But this is a one-person Croatian business whose blog stopped in February 2024. Judge the product on its transparency, and judge the purchase on whether the service is actually running the day you subscribe.
- Choosing a selection results in a full page refresh.
- Opens in a new window.