Umanitek Brings Guardian to ChatGPT and Claude
The new integration lets users investigate digital threats, create evidence packs and initiate takedown requests directly through their AI assistants.
Umanitek unveiled a new integration connecting its Guardian security platform with Claude and ChatGPT during its October 22 product webinar, demonstrating how users can manage digital threats through natural-language instructions.
Guardian × Claude: See how users can identify threats, create evidence packs, and initiate takedown requests directly through Claude.
From Conversation to Action
During the live demonstration, Umanitek co-founder Tomaz Levak used Claude to investigate accounts impersonating a protected entity on YouTube.
Through the Model Context Protocol (MCP), Claude retrieved threats identified by Guardian and organized using the OriginTrail Decentralized Knowledge Graph. Claude presented the findings directly in the conversation and generated a breakdown of harmful content across platforms.
Levak then instructed Claude to organize impersonating content into evidence packs. The resulting packs appeared in Guardian's dashboard without requiring him to open the platform.
When asked to initiate takedowns, Claude prompted Levak to specify which accounts to include before submitting the requests through Guardian.
The demonstration showed how users can move from investigating a threat to initiating action within a single AI conversation.
Guardian's Security Capabilities, Now Accessible Through AI
Guardian provides digital protection for individuals, brands and organizations through three core functions: threat detection, evidence collection and enforcement.
Depending on the subscription tier, Guardian monitors platforms including Facebook, Instagram, LinkedIn, TikTok, X, YouTube, Reddit, Telegram, Discord, Google and e-commerce platforms.
When threats are identified, users can create evidence packs that preserve relevant findings and supporting data. Guardian also supports takedown requests to remove infringing content.
The new integration makes these capabilities accessible through compatible AI assistants.
As Levak explained during the webinar, Guardian remains responsible for identifying threats. Claude and ChatGPT provide an additional interface through which users can access Guardian's intelligence and initiate actions.
Users can also draw on their AI assistant's existing capabilities to analyze findings and generate reports.
One Guardian Account, Multiple AI Assistants
The integration is built on MCP, an open protocol that allows AI assistants to connect with external tools and services.
This enables Guardian to work across compatible AI environments while maintaining a shared security context.
Levak explained that users can connect both Claude and ChatGPT to the same Guardian account, accessing the same threat intelligence, evidence packs and account permissions.
For example, evidence packs created through Claude can subsequently be accessed through ChatGPT, with actions synchronized through Guardian using the OriginTrail Decentralized Knowledge Graph.
The integration extends Guardian beyond its own dashboard, allowing users to work with the security platform through the AI assistants they already use.
Coming This Week
Umanitek plans to release the Guardian MCP integration during the week of October 26, with availability planned through Anthropic's and OpenAI's plugin catalogs.
The integration will be available across Guardian subscription tiers, with support for additional MCP-compatible assistants and agents planned.
The full demonstration is available in Umanitek's October 22 product webinar.