AI Tools Blocklist
Home AI Tools Database Taxonomy Pricing
Solutions
Enterprise IT & CISO Education Firewall Admins Shadow AI Prevention REST API Customer Login
Download Free Sample
Threat Intelligence

AI Tools Threat Intelligence Feed

Consume 17,410+ AI-tool domains as machine-readable intelligence. Daily exports in CSV, JSON, EDL, PAC, hosts, and DNS RPZ formats plus a REST API — every domain tagged with a primary category and multi-label subcategories your SOC can map to its own risk tiers.

17,410+AI Tool Domains
7Delivery Formats
DailyUpdate Cycle
18Categories
CSV / JSON / EDL Exports
STIX-Ready Structure
TIP Connectors
SIEM Enrichment
ATT&CK Guidance
Daily Updates
The Challenge

Why AI Tools Require Dedicated Threat Intelligence

Traditional threat feeds track malware C2 servers, phishing domains, and botnets. They are blind to the thousands of legitimate AI SaaS apps employees use daily to generate text, write code, create images, and analyze data.

The Classification Gap Your SIEM sees an HTTPS connection to an unknown domain. Your SOC analyst finds a clean reputation score and valid TLS certificate — and closes the ticket. The data is already gone.

Invisible Exfiltration Surface

AI tools are legitimate SaaS apps — not malicious. But pasting source code, contracts, or PII into them sends data through channels your threat stack doesn't track.

Detection Is Not the Problem — Classification Is

Your TIP doesn't know a domain is an AI tool or what category it falls into — the context you need to judge what it means for your data-protection requirements.

How Our Feed Solves This

We publish 17,410+ AI-tool domains as a structured, categorized domain list. Each record carries the classification context your SOC needs for triage.

Every record in the feed is
  • Classified into one of 18 categories (text generation, code assistants, image generation, voice synthesis, data analysis, and more)
  • Carrying a multi-label categories field ("Category > Subcategory") spanning 172 subcategories for granular policy decisions
  • Tagged with tool name, AI type, language, and active status — the context to map each category to your own risk tiers
  • Delivered as daily CSV, JSON, EDL, PAC, hosts, and DNS RPZ exports plus a REST API
  • Updated daily — built on a 102-million-domain pre-categorized corpus, with ~300,000 newly registered domains checked and classified every day
17,410+AI-tool domain indicators
102MPre-categorized domain corpus
300KNew domains checked daily
172Subcategories in the taxonomy
Machine-Readable Delivery

Structured Exports Built for Threat-Intel Pipelines

The feed is a plain, structured domain list — refreshed daily and published in CSV, JSON, EDL, PAC, hosts, and DNS RPZ formats plus a REST API. Every major TIP, SIEM, and SOAR platform can ingest at least one of these formats natively.

What Each Record Contains

Root Domain and Tool Name

The registrable domain of the AI tool, with aliases and subdomains resolved and dead domains pruned in the daily cycle.

Category Taxonomy

A primary category (one of 18) plus a multi-label categories field in "Category > Subcategory | ..." format spanning 172 subcategories.

AI Type, Language, and Status

Fields describing the kind of AI service, its language, and whether the domain is currently active — useful filters when building block policies.

Converting the List to STIX for Your TIP

If your TIP expects STIX bundles, a flat categorized domain list converts cleanly: each domain becomes one Indicator object, and our category fields become labels. You stay in control of the mapping.

You control the mapping
  • Encode the primary category and subcategories as STIX labels
  • Add your own risk-tier label per category (e.g. consumer chatbots = high, sanctioned enterprise assistants = low)
  • Diff against yesterday's export to import only newly added domains — no duplicates
# STIX 2.1 indicator generated from our CSV/JSON export
# by your own conversion script — you choose the labels and risk tier

{
  "type": "bundle",
  "id": "bundle--a3e2c8f1-9d4b-4e7a-b6c5-1f8d3e2a7b9c",
  "objects": [
    {
      "type": "indicator",
      "spec_version": "2.1",
      "id": "indicator--7c3a9f2e-1d5b-4e8c-a3f6-9b2d4e7c1a5f",
      "created": "2026-07-09T00:00:00.000Z",
      "modified": "2026-07-09T00:00:00.000Z",
      "name": "AI Text Generation Tool — example-ai-writer.com",
      "description": "Domain of an AI-powered text generation service. Primary category: text generation (from the AI Tools Blocklist).",
      "pattern": "[domain-name:value = 'example-ai-writer.com']",
      "pattern_type": "stix",
      "valid_from": "2026-07-09T00:00:00Z",
      "labels": ["ai-tool", "text-generation", "acme-tier-high"]
    }
  ]
}

# "ai-tool" and "text-generation" come straight from our category fields.
# "acme-tier-high" is the risk tier YOUR policy assigns to this category.

A Simple, Repeatable Conversion

Because the source data is a flat categorized list, the conversion is mechanical and easy to audit.

1

One Domain, One Indicator

Each root domain from the export becomes a STIX Indicator with a domain-name pattern.

2

Categories Become Labels

The primary category and subcategories map to STIX labels, so analysts see what kind of AI tool the domain is.

3

Your Policy Adds the Risk Context

Attach your own tier labels or ATT&CK tags per category — the guidance sections below suggest sensible defaults.

REST API

Authenticated REST endpoint returning the current list as structured JSON — the same fields as the CSV export.

  • Ideal for scripted pipelines that refresh a TIP or SIEM lookup table on a daily schedule

Category Filtering

Cherry-pick categories relevant to your risk profile, or ingest all 18.

  • Diffing consecutive daily exports gives you delta-only ingestion — no redundant transfers or duplicate processing
Platform Connectors

Native Integration with Threat Intelligence Platforms

Because the feed ships as plain CSV, JSON, and flat domain-list formats, the most widely deployed TIPs can ingest it with their standard import mechanisms. Configure once, and AI-tool indicators flow on the daily cycle.

PlatformIntegration MethodKey CapabilitySetup Time
MISPCSV/freetext feed URLAuto-correlation with existing network observables< 15 min
OpenCTICSV feed / import connectorIndicators in the knowledge graph with category labels< 15 min
ThreatConnectCSV/JSON source importCategory-to-tag mapping for playbook triggering< 15 min
Anomali ThreatStreamFlat-file feed importAutomated indicator enrichment< 15 min
Recorded FutureREST API pullCustom intelligence card configurations< 15 min

MISP Details

  • Indicators import as MISP attributes of type "domain" within tagged events, with our category names applied as MISP tags
  • MISP's pull mechanism fetches new events on schedule; correlation rules surface cases where your traffic logs already show connections to newly classified AI domains
  • Set MISP threat levels per category to match your own policy

OpenCTI Details

  • Point OpenCTI's CSV feed connector at the daily export URL — domains import as domain-name observables/indicators with our categories as labels
  • If you tag categories with ATT&CK techniques at ingest (see guidance below), those tags render in OpenCTI's matrix view
  • Apply your own scoring in OpenCTI per category to drive prioritization
ThreatConnect Import the CSV/JSON export as a source with tag-mapping rules. Category fields translate to ThreatConnect tags for automated playbook triggering; assign threat ratings per category to match your policy.
Automation

Automated Feed Ingestion and SIEM Enrichment

Many organizations run lean — a SIEM plus Python scripts that pull indicators and push them into lookup tables. Our REST API and daily JSON export are built for this workflow.

What the Script Below Does

1

Pull the Daily Export

Fetches the current AI-tool domain list as JSON from the REST API, authenticated with your API key.

2

Parse Records and Apply Your Tiers

Extracts domain, category, and subcategories from each record, then attaches the risk tier your policy assigns to that category.

3

Push to Splunk KV Store

Pushes enriched fields via Splunk REST API. Available as a lookup table in SPL queries. A cron job runs daily.

#!/usr/bin/env python3
# ai_feed_ingest.py — Pull the daily AI-tool domain export and push to Splunk

import os
import csv
import io
import requests

API_URL      = "https://www.aitoolsblocklist.com/api/database/?action=download_database"
API_KEY      = os.environ["AITBL_API_KEY"]
SPLUNK_URL   = "https://splunk.corp.local:8089"
SPLUNK_TOKEN = os.environ["SPLUNK_HEC_TOKEN"]

# YOUR policy: map feed categories to your own risk tiers
CATEGORY_TIER = {
    "AI Chatbots":         "high",
    "AI Writing Tools":    "high",
    "AI Code Assistants":  "high",
    "AI Image Generators": "medium",
    "AI Audio Tools":      "medium",
}

# Download the current database (CSV columns: domain,category,subcategory)
resp = requests.get(API_URL,
                    headers={"X-API-Key": API_KEY},
                    timeout=120)
resp.raise_for_status()
records = list(csv.DictReader(io.StringIO(resp.text)))

# Push each record to the Splunk KV store lookup table
for rec in records:
    category = rec.get("category", "unknown")
    entry = {
        "domain": rec["domain"],
        "category": category,
        "subcategory": rec.get("subcategory", ""),
        # risk tier comes from YOUR mapping, not from the feed
        "risk_tier": CATEGORY_TIER.get(category, "review"),
    }
    requests.post(
        f"{SPLUNK_URL}/servicesNS/nobody/search/storage/collections/data/ai_tool_iocs",
        headers={"Authorization": f"Bearer {SPLUNK_TOKEN}"},
        json=entry, verify=False
    )

print(f"Ingested {len(records)} AI-tool domains")

Adapt for Other Platforms

The core fetch-and-parse logic stays identical. Only the downstream push changes.

Elastic Security

Replace the Splunk KV push with an Elasticsearch bulk index operation targeting your threat-intel index.

Microsoft Sentinel

Use the Microsoft Graph Security tiIndicators API to push into Sentinel's TI data connector.

Chronicle (Google SecOps)

Use the Chronicle Ingestion API's UDM event format with indicator entity types.

Prefer No Code? Our REST API returns AI-tool records as a flat JSON array — domain, primary category, subcategories, AI type, language, and active status. Built for shell scripts, SOAR playbook HTTP actions, or SIEM modular inputs. Filter to the categories your policy cares about.
The Data Underneath

A Discovery Pipeline No Single SOC Could Replicate

Every indicator in the feed is backed by a continuously running classification engine — the same data, whichever of the seven delivery formats you consume.

17,410+AI-tool domains in the feed
102MPre-categorized domain corpus
300KNew domains checked every day
7Delivery formats (6 exports + API)
CSV JSON EDL PAC hosts DNS RPZ REST API
ATT&CK Framework

Mapping AI-Tool Categories to MITRE ATT&CK

The feed itself is a categorized domain list — it does not ship ATT&CK mappings. But the 18 categories map naturally onto specific ATT&CK techniques, and many teams tag indicators at ingest using guidance like the table below.

Suggested Category-to-Technique Mapping

Applying these category-level mappings lets your SOC operationalize AI-tool intelligence within existing ATT&CK-based detection frameworks, threat-hunting hypotheses, and risk-assessment methodologies.

AI CategoryPrimary TechniqueRisk Rationale
Text generation & language toolsT1567Employees paste sensitive text into web-based AI interfaces — data exfiltrated via legitimate HTTPS
Code assistantsT1059 + T1195.002Generate/execute code in production environments; AI-generated code may introduce vulnerabilities
Data analysis toolsT1530Integrate with cloud storage APIs (S3, GCS, Azure Blob) to process uploaded files containing regulated data
Voice & audio toolsT1123Process recorded audio that may contain sensitive conversations
T1567

Exfiltration Over Web Service

Primary exfiltration vector for all AI-tool categories. Sensitive data leaves through legitimate HTTPS connections not flagged by traditional DLP rules.

T1059

Command and Scripting Interpreter

Mapped to AI code assistants. These generate executable code employees may run in production without review — risking vulnerabilities or insecure configurations.

T1530

Data from Cloud Storage

Mapped to AI data-analysis tools connecting to cloud storage buckets (S3, GCS, Azure Blob). Risk: employees granting AI tools API access to regulated data.

T1048

Exfiltration Over Alternative Protocol

Mapped to AI tools using WebSocket, gRPC, or custom protocols. These may bypass traditional proxy inspection and require protocol-aware detection rules.

Closing the Detection Loop When a new AI-tool indicator arrives carrying the ATT&CK tags you applied at ingest, your TIP checks whether your detection stack covers those techniques. If coverage exists (e.g., a T1567 rule monitoring large outbound transfers), the indicator enriches that detection — the analyst knows the flagged connection is an upload to a known AI text-generation service. If coverage is missing, the indicator becomes a gap analysis data point and a reason to build that detection rule.
Risk Guidance

Category-Based Risk Tiering

Raw domain indicators are useful for blocking. Policy decisions require context. Every record ships with classification fields your team can map to its own risk tiers for automated triage.

Data Fields per Record
  • Root domain and tool name
  • Primary category (one of 18)
  • Multi-label categories field ("Category > Subcategory | ..."), 172 subcategories
  • AI type and language
  • Active status (dead domains pruned in the daily cycle)
  • Misclassification reported? Corrections ship in a subsequent daily export — email [email protected]

Risk Tier ≠ Reputation Score

We are not saying these domains are "malicious." Risk tiering is about potential for data exposure based on what a category of tools does with your data.

  • Typically higher risk — tools that process uploaded files, store conversation history, train on user inputs, or publish no privacy policy
  • Typically lower risk — tools that run locally in the browser, enterprise plans with SOC 2 compliance, data retention controls, no training on user data

Example Category-to-Tier Mapping

A starting point many customers adapt — map each of the 18 categories to a tier in your own policy.

CategoryExample Tier
Consumer chatbotsHigh — free-form paste of text and files
Code assistantsHigh for engineering orgs — source code exposure
Image/video generationMedium — less likely to receive regulated data
Sanctioned enterprise assistantsLow — allow-list the approved domains

Factors to Weigh When Assigning Tiers

Data Ingestion Method

Does the tool take text input, file uploads, API integrations, or screen capture?

Data Retention Policy

Review the vendor's published policy, if one exists.

Security Certifications

SOC 2, enterprise plans, compliance posture.

Jurisdiction

Consider data-residency implications for your regulatory scope.

Training on User Data

Whether the tool explicitly states it trains on inputs.

18Primary categories
172Subcategories
7Delivery formats (6 exports + API)
DailyUpdate cycle
Vendor Partnerships

Embedding AI-Tool Intelligence in Your Security Product

Building a CASB, SWG, DLP, SIEM, or SOAR product? Our feed provides a turnkey AI-tool classification layer you can embed directly — no need to build your own scanning and classification engine.

OEM Licensing Includes
  • Dedicated API endpoint — higher rate limits sized for vendor-scale ingestion
  • White-label feed exports — your vendor identity as the producer; customers see your product as the intelligence source
  • Custom category taxonomies — map our 18 categories to your product's URL category schema
  • Daily delta notifications — HTTP POST notification when each daily export publishes, listing the newly classified AI-tool domains — no polling required
  • Dedicated integration engineering — our team works with your developers to optimize the data pipeline
The Business Case Enterprise customers are asking every security vendor for AI-tool visibility. If your product can't identify AI domains, customers will supplement — or replace — it. Licensing our feed adds AI-tool classification in weeks, not months. You leverage a database tracking 17,410+ domains, built on a 102M-domain pre-categorized corpus and a daily discovery pipeline no individual vendor is likely to replicate.

White-Label Feeds

OEM feeds published under your vendor identity. Producer fields, feed names, and record descriptions carry your branding. Our infrastructure is invisible to your customers.

Daily Delta Notifications

HTTP POST when the daily export publishes, with the newly classified records in the payload — push alerts to customers without scheduling your own polling.

SOC Operations

SOC Workflow: From Indicator Ingestion to Incident Response

Ingesting indicators is step one. The real value comes from operationalizing them across the detection-response lifecycle.

1

Detection

Create SIEM correlation searches matching network traffic (proxy, DNS, firewall logs) against the AI-tool indicator lookup table.

  • Alerts enriched with category, your assigned risk tier, and any ATT&CK tags you applied at ingest before reaching the analyst
  • Categories you tier as high-risk → high-severity alerts; lower tiers → informational events for trend analysis
2

Triage

The analyst receives a context-rich alert — not a raw domain requiring manual research. What was a ten-minute investigation becomes a thirty-second triage decision.

"User jsmith connected to example-ai-writer.com — AI text generation tool — high-risk category under your AI policy — tagged T1567 (Exfiltration Over Web Service)"
3

Response

Based on alert context, the analyst can take graduated action:

  • Escalate to the user's manager
  • Create a DLP ticket to investigate what data was uploaded
  • Trigger a SOAR playbook: auto-block the domain, send Slack notification, open Jira ticket
  • For repeat offenders: auto-apply stricter web-filtering policies to that user's traffic
4

Reporting

Aggregate AI-tool alert data into weekly and monthly reports for CISO-level visibility. Our category taxonomy enables granular questions:

  • How many employees use AI code assistants vs. AI image generators?
  • Which departments generate the most AI-tool alerts?
  • Are AI-tool usage attempts trending up or down after policy changes?
  • Should specific tools be approved for specific teams, or should you invest in an enterprise AI platform?

Ready to Add AI-Tool Intelligence to Your Security Stack?

Download a free sample of the feed to test in your TIP today. Or tell us your platform and integration requirements — we will help you get ingestion running.

Request Your Threat Intelligence Feed

Tell us your TIP platform, preferred feed format (CSV, JSON, EDL, PAC, hosts, DNS RPZ, or REST API), and integration requirements — we will help you set up ingestion.

Keep Exploring

Related Resources