AI Tools Blocklist is a daily-updated database of 20,183+ AI-tool domains, organized into 18 functional categories and 165 subcategories. Plug it into your firewall, DNS filter, or proxy and take back control of the AI tools running on your network — in minutes, not months.
By Alpha Quantum — the team behind the 120M-domain Web Filtering Database.
Over 300 customers worldwide have trusted the AI services of our company — including
This blocklist is one slice of our Web Filtering Database — 120M+ domains classified into 59 filtering categories (adult, gambling, malware, streaming and more). Already licensing filtering data? Teams switching to us routinely pay less for equal or broader coverage.
See the Web Filtering DatabaseComplete coverage of the AI web — dead domains pruned daily, so you only ever block what’s live.
Allow coding copilots, block image generators. Granular enough to make policy — not an on/off switch.
Trains-on-your-data · data sovereignty · NSFW · deepfake/abusive — know which tools are dangerous, not just which exist.
New AI tools enter the list the day their domain is registered — each cohort re-checked 1–4 weeks later.
The list is distilled from screening ~99.99% of the active web — the same corpus behind our commercial Web Filtering Database.
hosts · EDL · PAC · SonicWall · DNSFilter · CSV/JSON · lookup API — it drops into whatever you already run.
Your employees and students are pasting source code, contracts, patient notes, and exam questions into AI tools you have never heard of. Every ungoverned prompt is a potential data leak — and the tools multiply faster than any manual list can track.
Nobody files a ticket before trying a new AI summarizer. Staff adopt tools the moment they trend, sign up with corporate email, and grant OAuth access to mailboxes and drives. By the time IT notices, sensitive data has already crossed the perimeter. Building a shadow-AI prevention program starts with knowing which domains to watch.
Text pasted into a chatbot may be retained, used for training, or exposed through provider breaches. For regulated data — GDPR personal records, HIPAA health information, FERPA student records, CIPA-covered school traffic — that retention alone can constitute a reportable incident. AI data-loss prevention begins at the DNS and firewall layer, before the prompt leaves the building.
Dozens of new AI tools launch every single day. A blocklist compiled last quarter — or scraped from a GitHub gist — misses the tools your users adopted this week. Freshness is not a nice-to-have; it is the entire product.
Compliance frameworks and cyber-insurance questionnaires increasingly demand evidence of AI-usage controls. "We block ChatGPT" is not an answer when thousands of alternatives exist. You need coverage you can document.
Blocking "all AI" breaks legitimate work and drives users to VPNs. The right answer is category-level policy: allow approved code assistants, block consumer chatbots, monitor the rest. That requires a taxonomy, not just a list.
Seven columns per domain: function and risk. These rows are copied verbatim from the free sample — download it and check them yourself.
| domain | primary_category | secondary_category | data_sovereignty_risk | abusive_purpose | trains_on_your_data |
|---|---|---|---|---|---|
| openai.com | Models & Infrastructure | Foundation models & model APIs | low | none | no |
| chatgpt.com | Text & Language | General assistants & chatbots | low | none | yes |
| deepseek.com | Models & Infrastructure | Foundation models & model APIs | high | none | yes |
| deeplivecam.net | Image & Visual | Face swap & effects | unknown | deepfake | unknown |
| crushon.ai | Lifestyle & Entertainment | AI companions & character chat | unknown | nsfw | yes |
| hackaigc.com | Text & Language | General assistants & chatbots | unknown | uncensored | unknown |
Note the difference on row two: openai.com (the API) doesn’t train on your data — chatgpt.com (the consumer app) does. That level of nuance is the product. Downloads also carry the full multi-label category list per domain.
The most complete inventory of AI-tool domains available anywhere — distilled from our main database of 120 million classified domains, not from someone’s bookmarks.
Re-exported every day. Roughly 300,000 newly registered domains are screened every 24 hours — so the tool that launched this morning is classified before your users find it.
Granular enough to allow “Code & Development” while blocking “Lifestyle & Entertainment.” Browse every branch on the taxonomy page.
Each domain lands in exactly one of 18 categories and one of 165 subcategories, so your policies stay unambiguous. Full definitions live at /taxonomy.php.
Autonomous agents, browser automation, workflow bots, and agentic orchestration platforms.
Voice cloning, text-to-speech, music generation, transcription, and audio enhancement tools.
Industry-specific AI for legal, finance, HR, real estate, healthcare, and other verticals.
Code assistants, autocomplete, AI IDEs, code review, and app-generation platforms.
AI helpdesks, support chatbots, ticket triage, and conversational service agents.
AI analytics, data-cleaning, research assistants, and automated insight generation.
AI design tools, 3D asset generation, UI mockups, and creative co-pilots.
Homework helpers, AI tutors, essay writers, and study aids — critical for schools.
Image generators, editors, upscalers, background removers, and visual style tools.
AI companions, character chat, games, astrology, and consumer entertainment apps.
Copy generators, SEO tools, ad-creative platforms, and sales outreach automation.
Foundation models, model APIs, inference hosting, fine-tuning, and vector databases.
AI note-takers, meeting assistants, document co-pilots, and team productivity tools.
Robotics platforms, embodied agents, autonomous systems, and simulation stacks.
AI search engines, answer engines, knowledge bases, and document Q&A tools.
AI-content detectors, deepfake detection, security co-pilots, and threat analysis.
Chatbots, writing assistants, translators, paraphrasers, and grammar tools.
Video generators, avatar creators, AI editing suites, and dubbing platforms.
Two pipelines feed one list: the 120M-domain corpus catches what exists, and a dedicated discovery pipeline catches what's new. Both run every day. Read the full write-up in our classification methodology.
Everything begins with our main pre-categorized database of 120 million domains — the same commercial database that powers websitecategorizationapi.com — where every domain is classified by function, content type, and risk category.
To keep the database current, roughly 300,000 newly registered domains from zone files, CT logs, and registration feeds are checked and classified every day. Many turn out to be empty or parked — those don't inflate the database.
Any domain in the corpus — existing or newly classified — that is identified as an AI tool automatically enters the AI-tools feed. This is how we catch tools that appear in no directory, no press coverage, and no curated list.
Classified AI-tool domains flow into the shared delivery step, where they merge with Pipeline B's output.
A dedicated pipeline monitors sources specifically frequented by AI tools: app directories, product-launch platforms, developer communities, academic repositories, and the open web — scanned daily, with overlapping sources by design.
Candidate domains are classified into the 18 categories with tool name and subcategory assignment. This catches tools that may not yet have enough web presence for the corpus classifiers — announced yesterday, in beta, or serving a niche vertical.
Dead domains are pruned, aliases and subdomains resolved — a clean, deduplicated set merges into the same daily cycle.
Both pipelines merge into one unified, deduplicated feed, exported daily to CSV, JSON, EDL, PAC, hosts, and RPZ, and served through the REST API — never more than 24 hours behind reality.
The same three columns power enterprise governance, classroom filtering, firewall rules, and commercial filtering products.
Turn "we should do something about AI" into an enforceable, auditable policy. Feed the database into your secure web gateway, CASB, or proxy to allow sanctioned tools, block risky categories, and generate the AI-usage evidence your auditors and cyber-insurers now request. Category granularity means you can approve your engineering team's code assistants while shutting down consumer chatbots company-wide.
Enterprise AI blockingProtect academic integrity and meet CIPA and FERPA obligations without playing whack-a-mole against every new homework-helper site. The Education & Learning and Text & Language categories cover essay writers, AI tutors, and paraphrasers students actually use — updated daily, because the site that circulated in the group chat this morning was not on last semester's list.
Education AI filteringSkip the parsing scripts. Pull a ready-made EDL straight into Palo Alto or Fortinet, load the RPZ zone into BIND, push the PAC file through group policy, or drop the hosts file onto managed endpoints. One cron job keeps every enforcement point synced with the daily export — no format wrangling, no stale entries, no 2 a.m. list maintenance.
Firewall admin guideAdd an "AI Tools" category to your DNS filter, parental-control app, or secure browser without building a discovery pipeline from scratch. OEM licensing gives you redistribution rights, the full 18-category taxonomy to expose in your own UI, and a daily feed your product can white-label. Two hundred-plus organizations — including filtering vendors — already build on this data.
OEM domain feedNo converters, no custom tooling. Full reference and endpoints in the API documentation.
Pull your plan's export, check when it last changed, and automate your daily sync with your personal API key.
# pull today's full export curl -H "X-API-Key: $KEY" \ "https://www.aitoolsblocklist.com/api/database/?action=download_database"
The canonical exports — three columns, quoted values, UTF-8. Ideal for SIEM enrichment, audits, and internal tooling.
domain,category,subcategory claude.ai,"Text & Language","General assistants & chatbots" openai.com,"Models & Infrastructure","Foundation models & model APIs"
External Dynamic List for Palo Alto and Fortinet firewalls. Point your firewall at the hosted URL and it refreshes itself — zero manual imports.
Response Policy Zone for BIND and compatible resolvers. Enforce blocks at the resolver so every device on the network is covered, unmanaged ones included.
Proxy auto-config for browser-level policy via GPO/MDM, and a classic hosts file for endpoint-level blocking on servers, kiosks, and lab machines.
Hands-on walkthroughs for the most common deployment targets. Each guide uses the same classified domain feed; pick the method that matches your stack.
Block AI tools at the OS level on Windows, macOS and Linux. Covers scripted updates via cron and Task Scheduler, Group Policy distribution, and Pi-hole/AdGuard integration.
Browser-level AI blocking via proxy auto-config. Deploy through GPO or MDM with category-aware rules and WPAD auto-discovery.
Point your Palo Alto, Fortinet, or any EDL-compatible firewall at a hosted URL and it refreshes itself automatically.
Category-specific guide for blocking ChatGPT, Gemini, Claude and the thousands of lesser-known chatbot tools employees find before IT does.
Govern Copilot, Cursor, Codeium and other code-generation tools that can exfiltrate proprietary source code through IDE plugins and browser extensions.
Pick a plan on the pricing page and pay with PayPal or any credit/debit card. Activation is immediate.
Your subscriber account at /login.php holds your personal API key, download links, and profile — everything in one place.
Schedule a cron job or point your firewall's EDL/RPZ fetcher at your feed URL. The API docs include copy-paste examples.
Apply category-level rules at your firewall, DNS resolver, or proxy. Allow what's sanctioned, block what isn't, and log the rest for review.
A weekend of scraping gets you a list. It does not get you a maintained dataset. Here is what the do-it-yourself route actually costs.
| Capability | DIY list (gists, scrapers, manual research) | AI Tools Blocklist |
|---|---|---|
| Coverage | A few hundred well-known tools; the long tail — where the risk lives — is invisible. | 20,183+ domains derived from our main 120M-domain categorization database. |
| Freshness | Updated when someone remembers. New tools go unnoticed for weeks or months. | 300K new domains screened and classified every day; daily exports. |
| Categorization | Flat lists — all-or-nothing blocking that breaks legitimate work. | 18 categories and 165 subcategories for surgical, defensible policy. |
| Formats | One format; you write and maintain the converters for everything else. | CSV, JSON, API, EDL, PAC, hosts, and RPZ out of the box. |
| Quality control | No review process; false positives block business sites and erode trust in the filter. | Automated classification with daily re-exports; reported misclassifications are corrected in subsequent daily releases. |
| Ongoing cost | Engineer-hours every week, forever — usually the most expensive line item. | From $99/month, fully maintained. Cancel anytime. |
Every paid plan includes a subscriber account with your personal API key, download links, and profile page at /login.php. Full details on the pricing page.
Pay securely with PayPal or credit/debit card on both subscription plans.
Our Web Filtering Database classifies 120M+ domains across the full URL-filtering taxonomy — adult, gambling, malware, streaming and dozens more. Already licensing filtering data? Teams switching to us routinely pay less for equal or broader coverage.
See the Web Filtering DatabaseDownload the 500-domain free sample, run it against your own traffic logs, and count how many AI tools are already active on your network. Most teams find dozens they had never heard of. Questions? Write to [email protected] — we are Alpha Quantum, based in Munich, Germany.
Tell us your requirements — firewall format, category filters, update frequency — and we will build a tailored feed.