Posture, finding, and alert summaries, read-only
AI Connectors Directory · Security
AI connectors for Security
17 connectors and apps are filed under Security across the ChatGPT and Claude directories, 0.5% of everything listed. All of them are on this page, with where each one can be reached and what security vendors, compliance platforms, and identity providers typically expose to an assistant.
17 connectors filed here
17 listed in ChatGPT
0 listed in Claude
0 on both marketplaces
The landscape
What Security looks like inside the assistants
At 17 connectors it is the 16th largest of 17 buckets, 0.5% of the directory. That is thin enough that the whole bucket fits on one page and a new entrant is not competing with hundreds of neighbours.
Of those, 17 listed in the ChatGPT apps directory. None of them is listed in both marketplaces yet.
One of the smallest buckets, and conspicuously conservative: most listings describe themselves as read-only and route anything consequential back to the vendor’s authenticated console, which is the right pattern for the category.
What to build
What security vendors, compliance platforms, and identity providers put in front of an assistant
Compliance framework and control explanations
Access request and approval workflows
Asset and identity inventory lookup
The same MCP server serves ChatGPT, Claude, and Microsoft Copilot. You build the tools, the authentication, and the permission model once, then decide which surfaces to distribute to. See where an MCP server can be distributed for how the platforms differ.
The full list
All 17 Security connectors
Listed alphabetically, as captured in August 2026. Each profile carries the marketplace listings, the publisher, and, where the endpoint answered, what the server actually exposes.
AccessOwl
Internal SaaS app catalogue and access requests
ChatGPT
AJAXX Data Scrubber
Read-only explainer for data broker exposure
ChatGPT
Bastion
Compliance posture and evidence review inside chat
ChatGPT
Bitdefender
Free URL check for malware and phishing
ChatGPT
Codex Security
Security scanning workflows for code and diffs
ChatGPT
McAfee
Scam detection for messages and links in chat
ChatGPT
Neura Relay MCP
Review layer for proposed agent actions
ChatGPT
PrivacyHawk
Data deletion and opt-out requests across brokers
ChatGPT
ProofX
Cryptographic proof of authorship for written work
ChatGPT
PureVPN Privacy Assistant
VPN server picks and connection troubleshooting
ChatGPT
Radar Lite
Domain email authentication and DNS security checks
ChatGPT
Rebound Scam Assistant
Risk assessment for suspicious messages
ChatGPT
SmartphoneKey
Resident management for smart locks
ChatGPT
Soluvery
Sharing audit for Google Drive files
ChatGPT
Sprinto
Compliance answers drawn from a Sprinto workspace
ChatGPT
ToolCheck by M8ven
Trust scoring for MCP servers and packages
ChatGPT
Vanta
Compliance status and fix drafting by request
ChatGPT
Reading this data
What the category column does and does not tell you
Categories come from the marketplace listings themselves, mapped onto one shared set, and at row level they are noisy: connectors are filed by their publisher, and plenty sit under a heading that does not describe the product well. Treat the grouping as a way to navigate the directory, not as a market segmentation.
The solid facts on these pages are the marketplace listings themselves (where each product is listed, what it declares, and its verification tier) plus the tool counts Node8 measured directly against live endpoints. The comparisons built from them are Node8's.
FAQ
Security connector questions, answered
How many AI connectors are there for Security?
17 connectors and apps were filed under Security across the ChatGPT and Claude marketplace directories as of August 2026, 0.5% of the 3,150 products in this directory. Every one of them is on this page.
What does a connector in the Security category actually do?
It connects a company's existing system to an AI assistant through an MCP server, so a user can ask about that company's data or trigger an action in conversation. In this category that usually means posture, finding, and alert summaries, read-only, and compliance framework and control explanations.
Does the same connector work in ChatGPT, Claude, and Copilot?
Largely, yes. MCP is a shared protocol, so one well-built server can serve ChatGPT, Claude, Microsoft Copilot, and other MCP clients. What differs is the listing and review process for each platform, plus some presentation details. Building for one and distributing to the rest is the usual sequence.
How long does it take to build and list one?
Most first connectors take eight to twelve weeks from kickoff to a live listing. The build itself is rarely the long pole. Deciding which data to expose, getting the underlying APIs into shape, and passing security and platform review are what set the timeline.
What does it cost?
Cost tracks scope: how many systems the connector reaches, how much work the underlying APIs need, and what the security requirements are. A single-source, read-only connector is a materially smaller project than one that writes into production systems across several backends. Node8 scopes this in a discovery session before any build commitment.
Keep browsing
Other categories
Productivity (760)Finance (363)Other (351)Business Operations (347)Data & Analytics (260)Travel (232)Developer Tools (188)Education & Research (130)Healthcare & Life Sciences (101)Entertainment & Media (95)Sales & Marketing (94)Creative & Design (92)Communication (44)Commerce & Shopping (35)Legal (33)Nonprofit (8)
Put your data where the questions are being asked
Node8 builds the MCP server, secures it, and takes it through review, so your data is reachable inside ChatGPT, Claude, and Copilot. Most first connectors ship in eight to twelve weeks.