AI Tools Category: Generative AI and LLM Websites
A dedicated web filtering category for generative AI and large language model tools. It covers 16,000+ AI-tool domains in 18 subcategories, updated daily, so your product can allow, warn or block AI use with precision.
domain: example-writer.ai
category_1: AI / LLM Tools
ai_subcategory: Text & Language
category_2: Business
iab_v3: Technology & Computing
# your policy decides: allow, warn, isolate or block
What the AI / LLM Tools category includes
A site lands here when its main purpose is to let people use an AI system: type a prompt, upload a file, generate media or run an agent. Coverage is split into 18 subcategories so policies can be as broad or as narrow as the deployment needs.
Text & Language
Chat assistants, writing and rewriting tools, summarizers, paraphrasers and AI translation services.
Image & Visual
Text-to-image generators, photo editors with generative fill, avatar makers and upscalers.
Video
Text-to-video, AI video editing, talking-head presenters, dubbing and clip generators.
Audio, Voice & Music
Voice cloning, text-to-speech, transcription, music and sound generation tools.
Code & Development
Coding assistants, code review bots, app builders from prompts and AI test generators.
Models & Infrastructure
Model hosting, inference APIs, model hubs, fine-tuning platforms and AI developer consoles.
Agents & Automation
Autonomous agents, browser agents and workflow automation driven by language models.
Productivity & Collab
AI meeting notes, email drafting, note-taking and AI features sold as standalone workspace tools.
Design, 3D & Creative
Logo and layout generators, 3D model creation, AI presentation and graphic design tools.
Education & Learning
AI tutors, homework helpers, quiz generators and study assistants aimed at learners.
Search, Knowledge & Docs
AI answer engines, chat-with-PDF tools, research assistants and document question answering.
Customer Support
AI chatbot builders for websites, help desk assistants and support automation platforms.
Marketing, Sales & SEO
AI copywriting for ads, SEO content generators, sales outreach and lead research assistants.
Data, Analytics & Research
AI spreadsheet and data analysis tools, chart builders and research data assistants.
Lifestyle & Entertainment
AI companions, character chat, horoscope and dating helpers, AI games and fun generators.
Security & Detection
AI content detectors, deepfake checkers and AI-driven security analysis tools.
Business & Vertical
AI tools for one industry, such as legal drafting, medical scribing, real estate or finance.
Robotics & Embodied AI
Sites for AI-driven robots, embodied AI platforms and their control or developer portals.
What it excludes, and the neighbouring categories
The category is about using AI tools, not reading about AI. Keeping the boundary tight stops a block rule from taking out news sites, universities or ordinary software vendors.
In the category
- Sites where the main product is an AI tool you use in the browser or through an account
- AI apps whose web domain handles sign-in, prompts, uploads or generated output
- Model hosting, inference and AI developer platforms
- AI wrappers: small sites that put a new interface on someone else's model
- AI browser extensions' own web domains and their back-end endpoints, when identifiable
Not in the category
- News, blogs and newsletters about AI, which go to News/Media or Blogs/Personal Sites
- University AI research labs, which go to Education/Reference or Science
- Software companies with a minor AI feature, which keep their main category
- AI conferences and events, which go to Business or Computers & Technology
- Generic cloud hosting used by AI apps, which goes to CDN / Hosting / Infrastructure
Typical example site types
Most of the 16,000+ domains are not famous brands. They are small, fast-moving tools, and many launch on new domains every week. These are the shapes you meet most often.
General chat assistants
Large, well-known assistants and the many smaller sites that offer similar chat on top of a licensed model.
Chat-with-your-document tools
Upload a contract, report or spreadsheet and ask questions. High value for users, high leakage risk for employers.
Single-purpose generators
One-page tools for headlines, cover letters, essays, product descriptions or social posts.
Media studios
Image, video and voice tools with credit-based plans, often with community galleries of generated work.
Agent platforms
Tools that act on the user's behalf: browse, fill in forms, send email or run code in a sandbox.
Model APIs and hubs
Developer consoles and inference endpoints that apps and scripts call directly, not only through a browser.
Who blocks AI tools and who allows them
AI policy is rarely a single block or allow decision. Most networks need different actions per subcategory, per user group and sometimes per time of day.
Allow
Approved tools under a company contract, AI tutors chosen by a district, coding assistants for engineering teams.
Allow and log
Search and knowledge tools where usage reports matter more than blocking, for example in higher education.
Warn or isolate
Unapproved chat assistants and document tools, opened with a warning page or in read-only browser isolation.
Block
AI companions for minors, unvetted file-upload tools in regulated sectors, AI proxies that evade filters.
| Environment | Typical action | Main reason | Subcategories often treated differently |
|---|---|---|---|
| K-12 schools (CIPA) | Mixed | Academic integrity, age limits in tool terms, student data | Allow chosen Education & Learning tools; block Lifestyle & Entertainment companions |
| Universities | Mostly allow | Research and teaching use; policy set by course | Watch Agents & Automation on lab networks; monitor exam rooms |
| Enterprises | Allow approved | Data leakage, contracts, records rules | Allow sanctioned tools; warn or block Search, Knowledge & Docs uploads to unvetted sites |
| Regulated sectors | Block by default | Client confidentiality, health and financial data rules | Exceptions per team for vetted Business & Vertical tools |
| Guest Wi-Fi | Allow | Visitors expect open access; no company data on the network | Block only where AI proxies are used to reach other blocked content |
| Parental control | Age-based | Age ratings, companion apps, image generation | Block companions and image tools for younger profiles; allow homework helpers with time limits |
Schools often ask whether CIPA requires blocking AI. CIPA itself targets visual depictions that are obscene, child pornography or harmful to minors, so AI policy is usually a local acceptable-use decision. See K-12 school web filtering for how districts set this up.
Multi-category cases
A domain can carry several categories, ranked by how much of the site each one describes. That ranking is what keeps AI rules from breaking ordinary business tools.
AI tutor for students
A school can allow it by subcategory while still blocking general chat assistants.
Design suite with a generative feature
The main product is design software, so a rule on the primary category alone leaves it open.
AI chat that unlocks blocked sites
Tools that fetch pages for the user can work as a proxy, so both categories apply.
AI companion with mature themes
Parental control profiles block on any matching category, not only the first.
Document chat with storage
Data loss rules for uploads should fire on either category.
AI coding platform
Engineering teams get an exception while other departments follow the default.
How vendors use the AI category in their products
The category is delivered as downloadable files that run inside your own systems, with a lookup API included in licences. Here is how different product teams put it to work.
Firewalls and secure web gateways
Load AI domains as a category or as an external dynamic list (EDL) and set allow, warn or block per user group.
DNS filters
Answer AI lookups from a DNS RPZ zone or a hosts-style list, which reaches apps and devices without an agent. Read DNS filtering vs web filtering for the trade-offs.
DLP and CASB tools
Use the category to decide when to inspect uploads and prompts, and focus inspection on document and chat tools.
Browser isolation
Open unapproved AI tools in an isolated session with copy, paste and upload controls turned off.
School and parental filters
Expose subcategories as simple toggles, so a teacher or parent can allow tutors and block companions.
MSP dashboards and reports
Report which AI tools are in use across client networks, grouped by subcategory, to start a shadow AI review.
Need to check where a single site lands first? Use the domain category lookup. For product teams, the OEM licensing page explains how the database is embedded in a filter or gateway.
Why AI tools need daily updates
AI tools appear faster than almost any other kind of website. A weekly or monthly feed misses the newest wrappers, which are often the least vetted.
Daily delta files
New AI-tool domains arrive as daily deltas, so a filter picks them up without a full reload of the database.
Second and third looks
New domains are checked again after a week and after a month, because many launch as a blank page and add the tool later.
Measured accuracy
Regular manual validations measure accuracy, and validation reports are available to licence holders.
A starting policy by AI subcategory
This matrix is a starting point many product teams use as default presets. Every school, company and family will tune it, which is exactly why the subcategories exist.
| AI subcategory | K-12 default | Enterprise default | Why it differs |
|---|---|---|---|
| Text & Language | Block | Approved only | Essay writing in school; pasted company text at work |
| Image & Visual | Block | Warn | Image tools can produce unsuitable content for minors |
| Video | Block | Warn | Heavy bandwidth and deepfake concerns |
| Audio, Voice & Music | Block | Warn | Voice cloning can be misused for fraud |
| Code & Development | Allow for CS classes | Approved only | Source code is sensitive company data |
| Models & Infrastructure | Block | Engineering only | Direct API access bypasses tool-level controls |
| Agents & Automation | Block | Block by default | Agents act with the user's sessions and data |
| Productivity & Collab | Block | Approved only | Meeting notes capture confidential discussion |
| Design, 3D & Creative | Teacher choice | Allow | Lower data risk, useful for art and marketing |
| Education & Learning | Allow chosen tools | Allow | Districts pick vetted tutors and quiz tools |
| Search, Knowledge & Docs | Older grades | Warn on upload | Document upload is the main leakage path |
| Customer Support | Block | Allow | A business tool with little use in class |
| Marketing, Sales & SEO | Block | Allow for teams | Customer lists may be uploaded for outreach |
| Data, Analytics & Research | Older grades | Approved only | Spreadsheets often hold personal data |
| Lifestyle & Entertainment | Block | Block | Companions and character chat raise safeguarding issues |
| Security & Detection | Allow for staff | Allow | Teachers use AI detectors; security teams use analysis tools |
| Business & Vertical | Block | Per department | Legal or medical tools fit one team, not everyone |
| Robotics & Embodied AI | Robotics clubs | Allow | Low risk, mostly developer and hobby portals |
Data leakage checklist for AI tools
For enterprises, the AI category is first a data protection control. These are the questions security teams work through when they turn it on.
Policy questions
- Which AI tools does the company hold a contract for, and with which data terms?
- Which teams need code, data or document tools, and which do not?
- Should unapproved tools be blocked outright, or opened with a warning page?
- Who reviews requests for new tools, and how fast?
Technical controls
- Category rules on the gateway and DNS layer, so mobile and remote users are covered
- Upload inspection on document, chat and storage categories together
- Logs grouped by AI subcategory for monthly shadow AI reports
- Daily feed updates, since new AI wrappers launch every day
Teams that sell into business can read more in web filtering software for business.
AI / LLM Tools category FAQ
What is the AI / LLM Tools web filtering category?
It covers websites whose main purpose is to give users access to AI tools: chat assistants, generators, coding assistants, agents and model platforms. It holds 16,000+ AI-tool domains in 18 subcategories and is updated daily.
Does the category include news about AI?
No. News and commentary about AI goes to News/Media or Computers & Technology. A business with a small AI feature keeps its main category and may carry AI tools as a secondary category.
Can I block only some AI tools and allow others?
Yes. The 18 subcategories let you allow approved coding or education tools while blocking companions, unapproved chat assistants or document upload tools.
Does blocking the category stop data leakage?
It closes the obvious paths to unapproved tools. Pair it with upload inspection and an allow list of approved tools, since AI features also live inside other software.
Are AI wrapper sites covered, not only the big assistants?
Yes. Small sites that put their own interface on a licensed model make up a large share of the set. They change names and domains often, which is why the feed is updated every day.
Will blocking the category break ordinary software with AI features?
Not if rules use the primary category. Office, design and support software with built-in AI keep their main category, with AI / LLM Tools ranked lower, so you choose whether secondary matches count.
How often are new AI tools added?
Daily. New domains are delivered in daily delta files and checked again after a week and after a month.
Can I license the AI tools category for my own product?
Yes, for internal use or OEM embedding. Starter database tiers start from $999; enterprise and OEM licences are quoted per deployment. See pricing or browse all 59 web filtering categories.
Add a precise AI category to your filter
16,000+ AI-tool domains in 18 subcategories, inside the same file as 120 million categorized domains. Tell us about your product and deployment.