iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
The reliable way to tell is to read the companion’s current technical documentation and privacy policy. Look for an explicit statement that responses are generated by a large language model (LLM), a named model or provider, or a description of messages and memories being sent to a model service. A scripted chatbot normally classifies your intent and selects a prepared reply or article. An LLM-based companion generates new text, sometimes with a retrieval system or scripts around it. Conversation style alone cannot prove which architecture is running, and many products combine both approaches.
What is the difference between a scripted chatbot and an LLM companion?
A scripted chatbot chooses from responses written in advance, while an LLM companion generates language at run time. The distinction matters because it affects what the system can do, what data it may send to providers, how predictable its answers are, and how you should verify anything important.
The GIZ guide to chatbots describes a typical FAQ bot as an intent-classification system: it identifies what a person wants and returns a predefined answer or a relevant article. A generative chatbot instead uses an LLM to compose a new response. Generative systems can also retrieve passages from a curated knowledge base before writing an answer; this is retrieval-augmented generation (RAG), and it remains generative rather than purely scripted.
The Office of the Australian Information Commissioner (OAIC) distinguishes an AI model from an AI system. The model is the language engine; the application can add prompts, scripts, retrieval, memory, moderation, business rules, tools and storage. Therefore, “uses an LLM” describes one component, not every part of the companion or its data practices.
#1 Best Overall
- Powerful Turbo Fan:CAGIWIRU Electric Air Duster have reached speed up to 130000RPM,Can efficient remove dust and debris.With LED light and Three Adjustable Speed to meet with different cleaning tasks.
- With Attached USB C Fasting Charging cable.Charing 3 hours,Enjoy up to 240 minutes of use on the lowest setting, For third speed can use for 25mins.with four charging options to meet with your needs.
- Wide Application: attached 5 different nozzles,making it suitable for a variety of scenes, such as car,pet, pc, keyboards, and other electronic devices. It also serves for office and home clean duster.
- Handy Design & Portable:Cordless air duster electric nozzle is Light weight and small size, design and comfortable to hold and space saving, convenient to carry around. You can put it in your room or inside the car without taking up space.
- Economical &Reusable: Built durable materials with a long-lasting battery,our CAGIWIRU air duster it's a sustainable alternative to disposable air cans.
| What you are comparing | Scripted or FAQ chatbot | LLM-based generative companion | Hybrid companion |
|---|---|---|---|
| How a reply is produced | Intent is matched to a prepared sentence, flow or article. | An LLM predicts and writes text from the conversation and its instructions. | Rules may handle some intents while an LLM handles open-ended turns. |
| Variation in wording | Usually limited to approved variants or templates. | New wording can be produced for each turn. | Varies by feature and fallback path. |
| Use of a knowledge base | May display a fixed article or answer. | May retrieve documents and generate an answer from them (RAG). | Can route some questions to search or articles and others to generation. |
| Typical failure | “I didn’t understand” or an irrelevant menu/article. | A fluent but incorrect, incomplete or invented answer. | Either type of failure, depending on the component used. |
| What documentation should reveal | Decision trees, intents, canned responses or FAQ articles. | LLM/model or provider names and language about generated responses. | A description of routing, tools, models and fallback behavior. |
Can an AI companion use both scripts and a language model?
Yes. “LLM or scripted” is often an oversimplification. A product can use a model to understand a message, apply a scripted safety flow, retrieve a fact, and then ask an LLM to phrase the final response. Another feature, such as account support or a crisis escalation, may bypass generation entirely.
Why a model name is not the whole architecture
An app may call an external model only for selected messages, add its own system prompt and safety filters, store conversation memory separately, and use non-generative code for notifications or actions. A named provider is useful evidence that a model is involved, but it does not tell you whether every reply is generated, what context is sent, or how long data is retained.
How retrieval changes the answer
RAG supplies a model with passages from a controlled knowledge base before generation. The retrieved text can improve relevance, but the final wording is still produced by the model and can contain errors. A citation or article link therefore shows where context came from; it does not prove that the response itself was a prewritten script.
The Tool Desk
Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →How do I check an AI companion’s official documentation?
Use the service’s own technical, AI, help and privacy pages before relying on a behavioral guess. Follow this order so that you separate architecture evidence from marketing language.
- Find an AI or technical page. Search the site for “how it works,” “generative AI,” “language model,” “LLM,” “inference,” “agents,” or “model provider.” An explicit sentence that replies are generated, or a named model/provider, is stronger evidence than a claim such as “smart” or “AI-powered.” The GIZ definitions and the OAIC model-versus-system explanation provide useful terminology for interpreting what you find.
- Check whether the statement applies to the companion feature. A company may describe an LLM used in a support assistant while its social or role-play companion uses a different pipeline. Look for product names, feature scopes, regional limits and dates on the page.
- Read release notes or architecture details for routing. Phrases such as “fallback,” “intent,” “knowledge base,” “tools,” “memory,” “moderation” and “human handoff” indicate additional components around the model. If the documentation says only that the service uses AI, it does not establish a particular architecture.
- Open the privacy policy and data-processing notices. Confirm what happens to text, voice recordings, transcripts, uploaded files and stored memories. These documents often identify a third-party LLM provider even when the marketing page does not.
- Record the date and edition. Models and providers can change without the conversation experience changing visibly. Save the page date or version, and treat an old help article as historical rather than current proof.
What should I look for in the privacy policy?
The privacy policy can reveal the data path more clearly than the chat interface. Look for four separate details: what leaves the app, who receives it, how long copies remain, and whether the material can be used for training or abuse monitoring.
Rank #2
- Screen Cleaner is Your Screen’s New Bestfriend! - EVEO is proud to present a screen cleaner that’s perfectly suited for your TV. Meet your new screen cleaner, computer screen cleaner, laptop cleaner, iPad cleaner, Macbook,tv cleaner or any other electronic device. Our screen cleaner spray and wipe along with the included plush microfiber cloth, easily removes dust, fingerprints & other smudges on the screen’s surface. Clean effectively for a streak-free screen
- Gentle on your Screens - Our screen cleaner spray and wipes is compatible to be used even for the most sensitive LCD, LED, CRT, and OLED screens in proper use. Be at ease knowing that this product is compatible with all the major brands: LG, SONY, Samsung, Sharp, iPad, iPhone, Android screens, Kindle, PC monitors - we’ve got you covered! Can be used as LCD screen cleaner, tv cleaner, smart tv screen cleaner, and more
- Premium Super Soft Plush Microfiber Cloth - Included in this efficient screen cleaning kit. This screen cleaner spray and wipe has our signature plush microfiber cloth specially designed to comply with screens and monitors and leave them streak-free. Has been shown to effectively absorbs stubborn stains, zero streaks, and no fiber shedding. Always use this together with our screen cleaner for the best results
- Easy-to-Use Screen Cleaner Kit + microfiber cleaning cloth - The EVEO Screen Cleaning Kit will easily be your new favorite tool when it comes to taking care of the things you love most, quick and easy solution for a streak-free screen. Simply screw on the sprayer gun on the screen cleaner bottle. Spray the cleaning liquid onto the microfiber cloth, then wipe the screen until desired cleanliness & shine is achieved.
- Quality Screen Care Kit Worth Purchasing - Enjoy a convenient Screen Cleaner anytime, anywhere you need it. Your satisfaction is our top priority. We stand behind the quality of our products and that’s why, you shouldn’t be worrying at all. We manufacture the Best Screen Cleaner. and can be used for laptop cleaning kit, laptop screen cleaner, tv screen cleaner for smart tv. MacBook screen cleaner, monitor cleaner or screen cleaner spray for electronic devices with microfiber cleaning cloth
Messages, voice and transcripts
Check whether typed messages, voice notes, audio files and speech-to-text transcripts are sent to an external model provider. “We process your content” is less specific than a statement identifying a provider or explaining that content is transmitted to generate a response.
Memory and context
Some companions send selected long-term memories along with the current message. Determine whether you can view, edit or delete those memories and whether deletion also removes copies held by a provider. A policy that mentions conversation history but not “memory” may still permit relevant past content to be included in a prompt, so ask the service if the wording is unclear.
Retention and training
Distinguish the time the app stores your conversation from the time a model provider keeps a transient copy. Also look for separate statements about service improvement, human review, safety monitoring and model training; permission for one purpose does not automatically answer the others.
For a concrete example, the EMBL-EBI AI chatbot disclaimer explicitly says its chatbots use generative AI and LLMs. It also warns that replies can be incorrect, unrelated, incomplete, outdated or misleading, and explains that personal information may be shared with third-party LLM providers under its setup. That combination—architecture disclosure plus a data warning—is the kind of evidence to look for.
The Companion privacy page states that a third-party LLM provider is used to understand messages and draft responses. It describes transmission of messages, voice notes and relevant stored memory, along with a provider copy retained for a limited period for abuse monitoring. Those are that service’s published terms, not a universal rule for every companion.
Rank #3
- 【Large size】：Measuring 800 x 300 mm/ 31. 5×12 inches, which is wider and taller than others. Large mouse pad can perfectly fit your keyboard and mouse of any size, and there is also ample space for peripherals such as phones, tablets, etc.
- 【LED backlight mouse pad】：Blade Hawks RGB mouse pad provides 7 static modes and 7 dynamic modes will give you the options you are looking for. Power-off memory function remembers the last lighting mode you selected. The super glow fiber Chroma will give you a colorful gaming setup that you want.
- 【Anti-slip rubber base】：Made from Natural rubber, this gaming mousepad will firmly grip your desktop, allowing you to enjoy the ultimate precision in the games you are most passionate about. It also comes with a texture that enhances this ability even further.
- 【Smooth and waterproof surface】: Micro-textured cloth surface, for extremely precise mouse control and a smooth movement. When liquid splashes on the gaming mouse pad, it forms water droplets and slides down. Will not delay your work or gameplay.
- 【Plug and play】: This mouse and keyboard pad is powered by USB, plug and play, no driver required. One button control light modes. Press one time to change the lighting mode, press twice to change the brightness, press and hold about 3 seconds turn ON/OFF.
Does an “AI-powered” label prove that replies come from an LLM?
No. “AI-powered” is a broad marketing description. It could refer to intent classification, recommendation software, speech recognition, moderation, search or an LLM, and it does not tell you whether the model handles every conversation.
Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Clear out junk files and repair common Windows errors3Fix the driver behind crashes, sound loss and screen glitchesLook for a specific disclosure that identifies the interaction as AI-generated and explains recording or sharing. For example, ElevenLabs’ agent disclosure guidance tells developers to inform users that they are interacting with AI and that conversations may be recorded and shared with the service and third-party LLM providers. A disclosure like this establishes an AI interaction and possible data recipients, but you still need the product’s own technical and privacy terms to learn which model is used and when.
Can I tell from the way the companion talks?
You can gather clues, but a chat experiment cannot identify the backend with confidence. A flexible answer may be generated, yet a scripted system can contain many branches, templates and retrieved articles. Conversely, an LLM can be tightly constrained by scripts, prompts, tools and safety rules.
Useful, but inconclusive, tests
- Ask the same question in two different phrasings and compare whether the answer adapts.
- Refer to a detail introduced several turns earlier and see whether the system uses it correctly.
- Ask a follow-up that requires combining two facts rather than selecting a menu option.
- Request a response in a new format, such as a short list or a different language, and observe whether the system follows the instruction.
- Ask the companion to explain where a factual answer came from, then check whether any cited source actually supports it.
These tests show behavior at the interface, not the implementation. A scripted flow can be designed to pass them, and an LLM can be prevented from doing so. Treat the results as prompts to inspect documentation, never as proof.
How should I compare two AI companions?
Use the same questions for each service and record answers from current official pages. Compare documented capabilities and data handling, not which companion sounds more natural.
Free tools Windows power users keep installed
One-click scans. No signup required.
Rank #4
- 【Softer and More Comfortable】Vaydeer wrist rest has unique diamond pattern, which is the combination of softness and aesthetics. The materials of wrist rest are improved into higher quality memory foam and covered with silky smooth lycra. The computer wrist rest makes you as comfortable and cushiony as like rest your wrists on clouds.
- 【Ergonomic Wrist Saver】The wrist rests for keyboard and mouse comes with a 17.32×3.15×0.83 inch keyboard wrist pad and a 5.94×3.15×0.83 inch mouse wrist support. Based on ergonomic design, the unique concave shape is the perfect fit for your wrist joints. The wrist rest pad fits most computer keyboards and laptops, improve hand and wrist posture, release your wrist and arm stress.
- 【Non-Slip Rubber Bottom】Featuring an anti-skid silicone base on the bottom, this wrist keyboard support stays firmly in place on your desk, preventing the padding from sliding around, ensuring stable and consistent wrist support during extended computer sessions.
- 【Better Experience & Pain Relief】Our keyboard arm rest is beneficial to alleviate the soreness caused by direct contact and friction between your arm and a hard desk surface, reducing the risk of wrist fatigue or carpal tunnel. The soft texture of memory foam can evenly distribute the pressure around your wrists and provide good support with just enough give.
- 【Helpful in Multiple Scenarios】Whether you're working, studying, writing, typing, gaming, this keyboard and mouse rest combo is an essential accessory to add comfort and support to your hands and wrists. It’s also a great gift for men, women, family, friend, coworker, gamer, teacher, etc.
| Comparison question | Evidence to record | Why it matters |
|---|---|---|
| What generates the response? | Scripted flows, an LLM, retrieval, tools, or a stated hybrid. | Explains flexibility, predictability and likely failure modes. |
| Is a model or provider named? | Exact model/provider name, feature scope and publication date. | Lets you verify the claim and detect that different features use different systems. |
| Does the app use retrieval or external tools? | Knowledge-base, search, browser, calendar or other tool descriptions. | Shows what information can enter a prompt and where answers may originate. |
| What content leaves the service? | Messages, voice, transcripts, files, profile data and stored memory. | Defines the personal information exposed to processors. |
| How long is it retained? | App retention, provider retention and deletion exceptions. | Affects the risk of later access, review or breach. |
| Can content train models? | Opt-in, opt-out, prohibited use or no clear statement. | Training use is a separate decision from generating your reply. |
| How are limitations disclosed? | Warnings about errors, outdated information, recording and human review. | Clear warnings indicate what the service expects users to verify. |
What does a named model or provider actually establish?
It establishes that the application claims to use that model or provider for the stated feature and period. It does not prove that the model sees your entire conversation, that every response is generated by it, or that the provider stores data under the same terms as the app.
Read the surrounding qualifiers: “may use,” “for selected requests,” “depending on availability,” “processed by our providers,” and regional or subscription restrictions. Also check whether the model name refers to speech recognition, embeddings, moderation or text generation. Those are different components even when one company supplies them.
Microsoft’s Copilot transparency note illustrates the broader point: a deployed assistant combines models with other technologies and can make mistakes. Use a model name as one piece of architecture evidence, then examine prompts, tools, memory and safeguards described by the application.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.How reliable are answers from an LLM companion?
Fluency and confidence are not reliability tests. Generated text can be factually wrong, omit important context or sound certain when the model lacks evidence. The OpenAI explanation of whether ChatGPT tells the truth describes this general limitation, and EMBL-EBI gives a similar warning for its own chatbot.
Windows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallOutdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchFor medical, legal, financial, safety, identity or relationship decisions, verify the claim with an authoritative source or qualified professional. Ask for the source, inspect the original document, and check dates and geographic scope. If the companion cannot explain its evidence, treat the answer as an unverified suggestion rather than a fact.
Best Value
- 【No Drill Mounted】Cable management tray can be assembled in as little as 3 minutes or less-simply. Side clips easily clampe the edge of desk, pass through the holes on the sides, and secure excess cables with baskets. Desk cable management supports inward or outward installation to meet your needs in different scenarios, and provides you with great convenience when collecting and organizing wires.
- 【Rubber Pad Protects Desk】Wire organizer under desk built-in soft rubber pads on the clamp protect your desk from scratches, so you will not damage your furniture or office.
- 【Sturdy and Beautiful】Cord management under desk is made of high-quality carbon steel and holds about 15 lbs. The wire basket has 2 holes on each side of the tray to enable your cables to pass freely. Cable baskets are suitable for desk thickness use from 0.4" to 2.4".
- 【Space Saving & Keep Organized】Desk cable management tray holds all your power cords, outlet strips, USB hubs and computer transformers hidden from your feet, and the data cables will not enter the vacuum cleaner to keep your desk clean and tidy. Desk wire organizer is also suitable for use in the kitchen and outdoors.
- 【Reduce Safety Risks】Cable tray under desk keeps all plugs away from your kids, no more worry about tripping. Using 2 holes wire tray saves extra space, helps you easily cabling and wire protection. The mesh design is not easy to accumulate dust, but avoids safety accidents caused by messy wires.
What should I do when the documentation is unclear?
Unclear documentation is itself a meaningful result: you cannot establish whether the system is scripted, generative or hybrid from the available public evidence. You can still reduce risk and obtain better information.
- Ask support a precise question. Request the model or provider used for the companion, whether all messages go through it, and whether voice and memory are included.
- Ask for retention and training terms separately. “Is my data used to improve the service?” and “How long are provider copies kept?” require different answers.
- Limit sensitive disclosures. Do not paste passwords, financial credentials, identity documents, confidential work material or another person’s private information while the data path is unknown.
- Use deletion controls. Remove individual messages and stored memories where possible, then check whether the policy describes backups, legal holds or provider copies.
- Save the policy version. A screenshot or downloaded copy helps you identify when terms changed, especially if the service updates models silently.
- Base your decision on documented controls. Choose the service whose architecture, retention, training use and limitations are explained clearly enough for your risk tolerance; do not choose solely on conversational realism.
What is the quickest trustworthy decision rule?
Classify the evidence into three levels. Confirmed generative use requires an explicit LLM or generated-response statement tied to the companion feature, ideally supported by privacy details naming the data path. Confirmed scripted behavior requires documentation describing intents, fixed replies, decision trees or article selection without a generative component. Undetermined or hybrid is the correct label when the service uses broad AI marketing, omits architecture details, or describes several components without saying which one handles your conversation.
When in doubt, report what is documented rather than guessing: “The service says it sends messages to a third-party LLM provider,” “The FAQ describes intent-based replies,” or “The public pages do not establish whether this feature is generated.” That wording is more accurate than inferring the backend from a charming, repetitive or surprisingly flexible conversation.
Frequently Asked Questions
Can I ask the companion itself which language model it uses?
You can ask, but the reply is not independent evidence: the companion may be unable to inspect its deployment or may generate an inaccurate answer. Treat it as a lead, then verify the model and provider in the service’s current technical and privacy documentation.
Is retrieval-augmented generation the same as a scripted chatbot?
No. Retrieval selects source passages, while an LLM still composes the final response. A system can combine retrieval with fixed templates, so the presence of a knowledge base does not by itself identify the whole architecture.
What does it mean if a privacy policy names a provider but not a model?
It shows that some processing may occur through that provider, but it does not reveal the exact model, feature coverage or routing rules. Ask whether the provider handles text generation, speech, moderation or another function, and whether the choice can change over time.
Should I stop using a companion if its architecture is not disclosed?
Not necessarily, but treat the unknown data path as a privacy and reliability risk. Avoid sensitive content, use deletion and opt-out controls, and prefer a service that clearly documents providers, retention, training use and limitations.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
