External / Web
SurfaceLooks for
Internet-facing web apps, APIs and TLS, mapped and fuzzed for weak spots.
Proves
Injection, authentication bypass and misconfiguration, proven on the target.
CAIVA.IO
Loading CAIVA.IO
CAIVA finds every asset you own without agents, shows which ones an attacker can reach, and proves it. A person approves before anything runs.
In a sample estate of 2,481 assets (72% running endpoint protection), CAIVA proved 17 attack paths an outsider could use. Four reach the finance database; fixing one server, dc-01, breaks three of them.
seenexploitableyou decideColour means the same thing on every chart.
Chapter 1 of 4: discover
CAIVA sweeps your network without agents or credentials. Servers, cloud accounts, network gear, databases and endpoints land on one graph, in cyan, the moment they are seen.
Chapter 2 of 4: reach
Most of the estate is noise. One route is not: from the internet to edge-proxy, then app-api-01, then dc-01, and on to your finance database, fin-db-02.
4 chains reach fin-db-02. 3 of them pass through dc-01.
Chapter 3 of 4: approve
CAIVA stops at the gate. Nothing is exploited until your analyst reviews the plan and approves it, and the proof stays inside the targets they signed off.
plan: app-api-01 to dc-01 to fin-db-02, bounded proof
Waiting for your approvalApproved by your analyst
Chapter 4 of 4: proven and fixed
Evidence comes back with every step logged. The last open chain runs through ci-runner, and that is your next fix.
A scanner tells you what might be wrong. CAIVA maps everything you own, ranks it by real-world risk, and proves what an attacker can actually exploit, with a person approving every step.
Too many raw scanner findings, ranked by CVSS alone.
One register, re-ranked by CISA KEV, EPSS and real exposure.
Unknown and shadow assets you were never counting.
Agentless discovery finds every device, managed or not.
“Is this actually exploitable, or just theoretically?”
A human-gated proof: bounded exploitation your analyst approves.
An audit needs evidence, not a story.
Evidence returned per finding, plus a full audit log of who did what.
Fixes still have to follow the change process.
Help Desk change approvals route remediation through your CAB.
Findings pile up and never close the loop.
SLA-tracked tickets that close only after a retest verifies the fix.
Every module, in flow order
It starts where your estate does — discovery and inventory — then risk, proof and remediation. Each box opens its own page with the real dashboards, and they share one inventory, so a finding in one is a ticket in another.
A credential-free sweep finds every device you own — managed, unmanaged or hiding behind NAT — and makes each a tracked asset.
Open in depthOne record per asset, with CIS benchmarks and licences reconciled.
Open in depthEvery asset scored by real-world risk, rolled up into one posture.
Open in depthOne CTEM register ranked by KEV, EPSS and exposure, with choke-points, departments and SLA tiers.
Open in depthFinds the reachable path across six isolated lanes and stops at a human gate — exploiting only after your analyst approves, then returns evidence.
Open in depthIncidents, problems, changes and requests, with SLAs, a CMDB and an end-user portal.
Open in depthBranded per-run pentest and vulnerability reports from the live estate.
Open in depthOne risk score for the whole estate, and whether it is improving.
Open in depthSSO, roles, teams, scanners, connectors and the audit log.
Open in depthSeven asset lanes orbit one core. Drag to orbit, or pick a lane to see what CAIVA knows about it.
The whole platform
Every area reads the same inventory and writes back to it. Drag to orbit, or pick an area to light its sub-modules and the data it exchanges.
IT Asset
Discover, inventory, score
The spine: a credential-free sweep finds every device, and each becomes one record that everything else reads.
Cybersecurity
Vulnerabilities · AI Pentest
The register ranked by real-world risk, and the pentest that proves what is exploitable at a human gate.
Help Desk
Native ITSM
Findings close as SLA-tracked tickets, with a CMDB synced from the inventory and change approvals on record.
Reports
Branded, from live data
Per-run pentest and vulnerability reports, and the board posture report, all built from the live estate.
Performance
Executive posture
One risk score for the estate and whether it is improving.
Administration
Access, identity, audit
Who can do what, how they sign in, what connects, and an append-only record of every action.
From every asset to one fix
CAIVA starts from everything you own and narrows to the one change that closes the most risk. Scroll to stack the layers; hover one to light it and its card.
AWS and Azure resolved without agents: instances, buckets and roles, each with its exposure and misconfigurations.
How CAIVA works
Seven stages, one estate. Each opens into its own module with the full dashboards.
A credential-free sweep resolves every host by IP, MAC and vendor, including the ones behind the firewall.
Discovery, imports and connectors resolve into one asset record, benchmarked against CIS, with licences reconciled.
KEV, EPSS, exposure and criticality score every asset and roll up into one posture you can track.
One CTEM register re-ranks raw CVSS by known exploitation and reach, then routes each finding to an SLA tier.
CAIVA runs isolated lanes and lights up the reachable path to the crown jewel. Exploitation waits for a person to approve, and it returns evidence, not a foothold.
Every run produces branded pentest and vulnerability reports from the live estate, so the board sees the trend.
Proven risk becomes an SLA-tracked ticket with an owner, and it closes only after a retest.
AI Pentest · lanes
Each asset type runs in its own sealed lane with its own depth posture — never a single engine pointed at everything. Some lanes stay at the surface; others go deep only once they have credentials.
Looks for
Internet-facing web apps, APIs and TLS, mapped and fuzzed for weak spots.
Proves
Injection, authentication bypass and misconfiguration, proven on the target.
Looks for
Open services, host exposure and weak configuration across the network.
Proves
Remote code execution and credential attacks, bounded to the approved scope.
Looks for
Domain layout, trusts and the reachable attack paths through it.
Proves
Privilege escalation toward domain control, step by step.
AWS · Azure · Entra
Looks for
Cloud and identity misconfiguration across your accounts.
Proves
Privilege escalation and lateral movement between accounts.
Looks for
Cluster, workload and image posture.
Proves
Container escape and lateral movement inside the cluster.
Detection-only — secrets & code
Looks for
Exposed secrets, risky dependencies and insecure code.
Proves
no exploitation in this lane — findings only
Every lane stops at the same human gate before anything is exploited, and returns evidence, not a foothold.
Open AI PentestCause chain
The hero showed the path. This is the record behind it: every step, the module that took it, and the evidence it left. Play it, scroll it, or pick any step.
Asset Discovery
edge-proxy · 203.0.113.24:443
Verdict
Follow the chain to the verdict.
The fourth chain, via ci-runner, stays open and tracked.
Sample run: IDs and times are illustrative. CVSS is the NVD base severity score out of 10; EPSS is the probability, 0 to 1, that a CVE is exploited in the next 30 days; KEV means CISA lists it as exploited in the wild.
Inside the operation
Scroll to walk the floor: discovery, triage, the human gate, then the proven path on the big board.
Business use cases
Pick a question to watch it travel through CAIVA to an answer. Every step is a module you can open.
A KEV-listed CVE hits the news. Are we exposed?
Outcome: Answered from the live inventory: three affected, one reachable, one 7-day ticket.
Built for a bank’s controls
Nine controls CAIVA enforces on every run and every fix. Each one opens the screen where it lives.
Regulated banks first, across cloud and on-premises estates, and the roles that carry the risk day to day — each with what they do in CAIVA.
Board-ready posture and whether it is improving, in one score and its trend.
Open PerformanceTriage the register by real risk, known-exploited and reachable first, and route each fix to its owner.
Open VulnerabilitiesOne live inventory and CMDB, with findings as SLA-tracked tickets.
Open Help DeskEvidence per finding, the audit log, and who approved what, on demand.
Open AdministrationRequest a demo
Tell us about your environment. We’ll show you the path an attacker would take through it, the gate your analyst approves, the proof, and the one fix that closes the most.