Integrate with us
Fit the platform into the tools you already run.
Everything the console does is available over a documented API, and the results are built to flow into your pipeline, your GRC tool and your auditor's workflow.
Integration surfaces
Eleven ways in.
Assistants
ChatGPT & Claude
Ask about your security posture in plain language, and check an MCP server before you install it. One MCP server with OAuth sign-in, read-only by default; your role's permissions apply.
Supply chain
MCP servers you install
Check every server in your Claude, Cursor or VS Code config against signed scans, fail a pull request that adds one needing review, and gate launches in ToolHive.
Core
REST API
Start scans, read findings and the estate graph, manage assets and schedules. Token-authenticated and documented endpoint by endpoint.
Build
CI/CD & pull requests
A GitHub Action gates the build on proven risk. Fixes arrive as draft pull requests against your own repository.
Compliance
GRC & compliance
Every report ships a machine-readable twin so evidence can flow into Vanta, Drata or Secureframe.
Cloud
Cloud accounts
Agentless connection to AWS, Azure and GCP — Terraform for AWS, an app registration for Azure, a service account for GCP. Credentials are verified with a real call before they are stored.
Events
Signed webhooks
Trust-state changes for your agents are emitted as signed, time-limited events you can verify offline and deliver to your own endpoint.
Verify
Agent-identity verification
A site or service that receives a request from an agent can check a signed receipt without an account, and get a covered / not-covered answer.
Runtime
Runtime & endpoint
An optional detection-only runtime sensor, and a connection to a third-party prevention engine for any blocking action.
Network
Internal networks
A thin outbound-only connector lets us test networks that are not reachable from the internet, under signed authorization.
AI assistants and MCP
Where you can use Redthread today.
The Redthread connector is one MCP server at https://app.redthreadsec.com/mcp with OAuth sign-in. You can add it by that address now; the directory listings make it one click once they are approved.
| Where | How | Status |
|---|---|---|
| Claude (web, desktop, mobile, Claude Code) | Settings → Connectors → Add custom connector, with the address above | Available now |
| Claude connector directory | One-click listing for Claude users | In review |
| ChatGPT | Settings → Apps & Connectors → create a connector (developer mode, or added by a workspace admin) | Available now |
| ChatGPT app directory | One-click listing for ChatGPT users | In review |
| Official MCP Registry | Listed as com.redthreadsec/redthread, for any MCP client that reads the registry | Available now |
| Any MCP client with OAuth | Point it at the address above | Available now |
| GitHub Actions | Scan and gate on proven risk; check-config fails a pull request that adds an MCP server needing review | Available now |
| ToolHive | Launch gate and a permission profile from your own policy | Preview |
Status as of 12 October 2026. Step-by-step setup: the ChatGPT & Claude guide.
For integrators
Keep the evidence intact.
Building an integration?
Start with the API reference. Tell us what you are building and we will help you scope it, and make sure the evidence labels carry through to your side.
Ground rules
- Integrations never get write access to a customer's cloud.
- Findings keep their evidence label — observed, reported or proven — wherever they are shown.
- Anything that blocks runs through a connected partner engine, not our sensor.
Integrate with Redthread.
Tell us about your product and the workflow you want to connect.