DEVELOPER PLATFORM
One platform. Five surfaces. The same forensic engine that powers your 4D Diagnostic — now reachable from your code, your agents, and your pipelines.
Every byte of your 4D forensic data — security findings, cost leaks, architectural drift, compliance posture — flows through the same authoritative store. Pick the surface that matches how you build.
api.cloudsealed.com
JSON over HTTPS. Pull findings, list resources, trigger reanalysis, and power custom dashboards.
Read the referencemcp.cloudsealed.com
Drop CloudSealed into Claude, Cursor, or any MCP-compatible agent. Audit data becomes a tool call away.
Install the serverskill bundle
A packaged Skill for Claude that knows how to read findings, draft remediations, and reason over your 4D posture.
Add the SkillTypeScript · Python · Go · C#
First-class clients with types, retries, and pagination baked in. Six lines from install to first finding.
Pick a languageHMAC-signed events
Get pushed the moment a critical finding lands. Wire CloudSealed into PagerDuty, Slack, Jira, or your own service.
Subscribe to eventsSame data model, three languages, one promise: no glue code between you and the forensic engine.
import { CloudSealed } from "@cloudsealed/sdk";
const cs = new CloudSealed({ token: process.env.CS_TOKEN });
const { data } = await cs.findings.list({ severity: "critical" });
console.log(data[0]);OAuth 2.0 with granular scopes — the same token works for REST, MCP, SDKs, and webhooks. Issue, rotate, and revoke from one console.
Every credential ships read-only. Write operations require an explicit scope upgrade with an audit trail. No accidental writes from a leaked PAT.
Every endpoint is documented in OpenAPI 3.1. SDKs are generated from the same spec. No drift between docs and runtime.
Tool schemas, deterministic error codes, and bounded responses — so an LLM can call CloudSealed without hallucinating shape or scope.
Spin up a sandbox tenant, mint a token, and pull your first finding in under five minutes.