ConnectionsNew + improved · included on every plan

Your stack, put to work.

Every connection checks its own access before you rely on it, and counts only toward the audits you choose. Then it takes on real jobs: tickets, vendor reviews, alerts and the fixes you approve.

Connect Oktastep 4 of 5 · check access0of 38 checks can run
okta.users.readCalling Okta…Checking
okta.factors.read
okta.policies.read
okta.apps.read
okta.logs.read
okta.iam.read
6 checks can’t run until Okta can see who holds admin roles.Add a custom admin role with only “View roles, resources, and admin assignments”. The token stays read-only.
From Settings › Connections. One live call per permission, before the first scan.
One connection, several jobs · turn them on and off
CONNECT YOUR IDP4 scopes, all read-only
readokta.users.read
readokta.factors.read
readokta.policies.read
readokta.groups.read
Each job asks only for the access it needs, and reports its own health. A broken vendor sync never hides behind a healthy scan.

Connected isn’t the same as covered.

The same Okta token, read two ways. Flip it.

ConnectedIntegration statusThe token is saved and the logo turns green. Nothing says what that token can’t read, so you find out at audit time.
Assumed coveredCan’t run · not evidenced, never passing
What’s different

A green tick is where most setups end. It’s where ours start.

01

“Supported” should mean it worked this morning.

Every catalog entry shows how many customer tenants it ran on this week, and when it last succeeded. New ones say they’re new. We run each connection ourselves every day, and when our own run fails, the entry says so before you find out.

Vanta lists more integrations than we do, 400 and counting. We’d rather show you which of ours worked today.

OktaChecks your policiesProven on 96 tenants · last success just now
GitHubChecks your policiesProven on 141 tenants · last success 17m ago
1PasswordChecks your policiesNew · not yet proven on a customer tenant
DatadogChecks your policiesProven on 41 tenants · our daily check failing since 9 Oct
SnykNeeds a paid Snyk planProven on 63 tenants · last success 31m ago
Tenant counts are illustrative. The live catalog reads them from the connector registry.
The catalog · cloud, identity, code, HR, procurement, security toolsNot listed? Describe it in a manifest →
AWScloudGoogle CloudcloudAzurecloudKubernetescloudOktaidentityEntra IDidentityGoogle WorkspaceidentityMicrosoft 365productivity1PasswordidentityGitHubcodeJiraworkLinearworkSlackalertsTeamsalertsPagerDutyincidentDatadogobservabilityAWScloudGoogle CloudcloudAzurecloudKubernetescloudOktaidentityEntra IDidentityGoogle WorkspaceidentityMicrosoft 365productivity1PasswordidentityGitHubcodeJiraworkLinearworkSlackalertsTeamsalertsPagerDutyincidentDatadogobservability
BambooHRHRRipplingHRCoupaprocurementZipprocurementJamfdevicesSnyksecurityWizsecurityProwlerimportCheckovimportTrivyimportScubaGearimportOSV-ScannerimportAny REST APImanifestYour scanner’s reportimportBambooHRHRRipplingHRCoupaprocurementZipprocurementJamfdevicesSnyksecurityWizsecurityProwlerimportCheckovimportTrivyimportScubaGearimportOSV-ScannerimportAny REST APImanifestYour scanner’s reportimport
02

Hear it broke the day it breaks. Not the day you need the evidence.

An expired token, a permission someone removed, a vendor plan that changed. The moment we can’t reach a platform, the connection’s owner gets an email naming what broke, which checks it stops, and the one thing to do. It also goes on a single Needs you list in the app.

Evidence collected before the break stays on the record. The gap is dated, so your auditor sees exactly how long a check was dark, and that someone was told.

Break something
Needs you1
Slack · token expired8 checks paused. Evidence up to 09:14 today is kept.Reconnect
Email to [email protected] · sent within minutesWe can’t reach SlackSlack’s token expired at 09:14. 8 checks are paused until it’s reconnected. Nothing collected before then is lost.
03

Connect everything. Only what’s in scope counts.

No auditor has asked about MFA on your on-call system, and none should see it fail on your SOC 2. The last step of setup asks which audits a connection counts toward. It’s the same scope your assessment uses, and every change is dated on the trail.

Connect PagerDuty · last step · try it
SOC 2 Type II · FY27Incident response is in scope
HIPAA · 2026PagerDuty doesn’t store or process PHI
ISO 27001 · 2027Scope not set yet. Decide later.
PagerDuty’s 8 results count toward SOC 2 Type II.
04

Not in the catalog isn’t a dead end.

Paste a key, grant a federated role, upload a scanner’s report, or describe any REST API in a short YAML manifest. Imported results are labelled as imported, and go stale if the next report doesn’t arrive.

Vanta’s Private Integrations let you push your own data in through its API, which is engineering work. A manifest is one file, and we dry-run it against your real vendor before it goes live.

freshservice.connector.yaml
Dry run passed · 12 checks against acme.freshservice.com
05

The tools you already pay for get more useful.

A new supplier in Coupa or Zip starts a vendor review here, with the questionnaire and evidence request already sent. When security decides, the decision goes back to the supplier record, so procurement never asks twice and never waits on a spreadsheet. Your procurement system now holds a security answer it didn’t have before.

Coupa · supplier recordNorthwind Analytics
Requested byLena Park · Marketing
Data sharedCustomer emails, usage events
Security statusPending security review
vCISO Lite · vendor review
Review openedStarted by the new supplier in Coupa. Nobody filed a request.
Questionnaire sentTo Northwind, with an evidence request for SOC 2 and a DPA
Approved by Maria ChenWith one condition: SSO before go-live
Written back to CoupaStatus, conditions and review date on the supplier record
06

Read-only until you say otherwise.

Every connection starts read-only. When you want fixes applied for you, an admin issues a write grant on that connection: scoped to the changes it allows, with an expiry and a reason on record.

Sprinto’s Autonomous Trust Platform remediates across your systems for you. Here, nothing writes without a grant you can see and revoke, and every fix it enables still closes only on a passing rescan.

Write grant · AWSActive
AllowsS3 public access blocks, security group rules
Issued byMaria Chen · “Q4 hardening sprint”
Expiresin 6d 11h 42m
In the appSettings → ConnectionsPlanIncluded on every plan. No per-connection fee.To startA read-only credential for one platform you already use

Related reading

Compliance

GRC Software: 2026 Buyer’s Guide

Compare GRC software for 2026: six platforms across the platform-augmented, mid-market and enterprise tiers, in one buyer’s view.

Read the guide →

Connect it once. Then put it to work.

Every connection proves what it can read, counts only where you say, and does the jobs you give it.