The security platform that explains itself.
Kandaro catches attackers who are already inside your network, then tells you in plain English what happened and what to do next. Built and hosted in the EU.
Deception · Overview
Good morning. 2 incidents need you.
Open incidents
2
Events
47
Decoys online
12/12
Bait planted
18
Finance workstation probed the backup NAS and tried to log in
Explained by Kandaro AI
At 02:14 a machine on the finance VLAN (10.20.4.77) scanned the decoy NAS, then tried to log in over SSH and SMB as backup-admin. Nobody has a reason to touch this device, so treat it as hands-on-keyboard activity.
- 02:21:40smb.share_enum · nas-backup-01high
- 02:19:12login.password · nas-backup-01high
- 02:17:55mcp.tools_call · ai-gatewaycritical
- 02:16:03ssh.auth · nas-backup-01medium
- 02:14:07tcp.port_scan · nas-backup-01medium
- 01:58:31snmp.get · printer-3flow
- 01:42:10honeycred.used · finance-sharecritical
- 01:39:48mysql.query · db-replicahigh
- 02:21:40smb.share_enum · nas-backup-01high
- 02:19:12login.password · nas-backup-01high
- 02:17:55mcp.tools_call · ai-gatewaycritical
- 02:16:03ssh.auth · nas-backup-01medium
- 02:14:07tcp.port_scan · nas-backup-01medium
- 01:58:31snmp.get · printer-3flow
- 01:42:10honeycred.used · finance-sharecritical
- 01:39:48mysql.query · db-replicahigh
Why Kandaro
Why security teams choose Kandaro.
Most security tools hand you more data. Kandaro closes the gap between an alert and a decision: it only speaks up when something real happens, and it tells you what that was and what to do about it.
Signal over noise
A decoy has no users and no purpose, so nobody has a reason to touch it. When something does, it is worth your attention, and worth your SOC’s. No tuning, no baselines, no alert fatigue.
Answers, not alerts
Every finding arrives as what happened, why it matters and what to do next, in plain English. The person on call at 2 a.m. can act on it without a specialist.
Ready for AI attackers
Attackers now send agents ahead of them. Kandaro runs decoy MCP servers and LLM APIs that only automated agents talk to, so you see them the moment they start exploring. Decoy tools among your own agents’ tools are planned next.
European engineering
Netherlands-based and privacy-first. Hosted in the EU, with the AI running on our own infrastructure by default, so data about your network stays here.
The products
We catch whoever gets in. Next, we attack you first.
Deception is live today: decoys that catch attackers who are already inside. Emulation is next: an AI adversary that tests your defenses continuously. Then come Investigation, an AI SOC that decides what each alert means, and Protection, which acts on it. One job each: detect, test, decide, act. All four run in the Kandaro Console, with the same AI explaining every finding.
Decoys that look like your real devices, and tell you who touched them.
NAS boxes, printers, servers and AI tool servers that nobody should ever use. When something does, you get an incident with a narrative instead of a packet capture.
- Decoys running in about a minute
- Bait credentials & breadcrumbs
- Tarpits with endless fake data
- Decoys for AI agents (MCP, LLM APIs)
Finance workstation probed the backup NAS and tried to log in
Explained by Kandaro AI
At 02:14 a machine on the finance VLAN (10.20.4.77) scanned the decoy NAS, then tried to log in over SSH and SMB as backup-admin. Nobody has a reason to touch this device, so treat it as hands-on-keyboard activity.
- nas-backup-01NAShigh
- ai-gatewayAI agent serverhigh
- printer-3fPrinter
- db-replicaLinux server
An AI adversary that attacks you before attackers do.
Continuous adversary emulation, run by AI. It will attack your environment the way real attackers do, and show what was blocked, what was detected and what got through.
- A threat library that stays current
- Continuous regression testing
- Automatic blocked / detected checks
- Evidence for DORA and NIS2
Access
Execute
Discover
Lateral
Creds
Exfil
An AI SOC that investigates every alert, not only ours.
It will take the alerts from your EDR, SIEM, identity and cloud tools together with Kandaro’s own, investigate each one, explain its verdict and recommend what to do.
- Triages EDR, SIEM, identity & cloud alerts
- Explains every verdict in plain language
- Recommends the next step
- Works on alerts from other tools, not only Kandaro’s
Alert queue · investigated
- Kandaro DeceptionDecoy NAS login from 10.20.4.77Escalated
- EDREncoded PowerShell on FIN-WS-114Escalated · same machine
- IdentitySign-in from a new country for j.devriesClosed · company VPN exit
- CloudStorage bucket policy changedNeeds a person
When something is caught, it acts.
The product that acts. When a decoy is touched or you approve a response, it will stop the agent session, isolate the machine or disable the account.
- Stops a hijacked AI agent’s session
- Isolates machines, disables accounts
- Alert, ask first or act: your choice
Agent session · support-assistant
- 01tickets.searchok
- 02docs.readok
- 03web.fetchok · page contained hidden instructions
- 04vault.read_secretdecoy tool
Session stopped. No real tool was reached after the decoy.
AI that explains
From raw protocol data to a sentence your whole team understands.
Security tools are fluent in packets. Most teams aren’t, especially at 2 a.m. Kandaro turns every finding into what happened, why it matters and what to do next.
{
"sensor": "nas-backup-01",
"profile": "nas",
"proto": "smb2",
"type": "login_attempt/ntlm",
"src": "10.20.4.77:51322",
"dst": "10.20.8.250:445",
"ntlm": {
"user": "backup-admin",
"domain": "ACME",
"workstation": "FIN-WS-114"
},
"result": "STATUS_LOGON_FAILURE",
"ts": "2026-10-08T02:19:12Z"
}Someone on the finance network tried to sign in to your backup NAS.
The workstation FIN-WS-114 (10.20.4.77) tried the account backup-admin on a decoy file share. This device isn't real, so no legitimate user or job should ever connect to it.
Recommended action
Isolate FIN-WS-114, then reset the backup-admin password and check where it is used.
- What happened
- A plain-English account of the activity, with the raw detail one click away.
- Why it matters
- Context on the device, the account and the pattern, so severity makes sense.
- What to do next
- A concrete recommended action, sent to email, Slack, your SIEM or your own systems.
The AI runs on Kandaro's own EU infrastructure by default. Attacker-supplied text is fenced off from its instructions, and a rule-based severity floor means the model can raise a severity, never lower it.
Kandaro Console
One cockpit. One continuous cycle.
Security is not a yearly project. From the Console you will pick attacks from the library, let Emulation run them, see what your defenses did and improve them, then go round again. Deception stays on the whole time and catches whoever gets through anyway. The Console comes with every product.
One view
Decoy hits, emulation results and open incidents on one overview, explained the same way.
One account, MFA required
Invite your team once. Every user signs in with a passkey or TOTP.
One alert pipeline
Email, Slack, SIEM connectors and signed webhooks for every product, filtered by severity.
The continuous cycle
Emulation Coming soon01
Select
Pick scenarios from the threat library, or build your own.
02
Attack
Emulation runs them against your own environment.
03
Verify
See what was blocked, detected and missed.
04
Improve
Close the gap, then re-run to prove the fix.
Repeats on your schedule
Deception LiveAlways on
Decoys watch the whole time and catch whoever gets through anyway, including through the gaps you have not closed yet.
One AI layer explains every finding, from every product
Kandaro Investigation
An AI SOC with its own ground truth.
Kandaro Investigation, in development, will investigate the alerts from your security tools. It will not work alone: Deception gives it alerts with no innocent explanation, and Emulation will check that it catches a real attack.
Deception feeds it Live
- Nobody has a reason to touch a decoy, so a Deception incident has no innocent explanation.
- Each incident carries MITRE ATT&CK techniques, a plain-English explanation and a recommended action.
- Investigation will use these as fixed points when it weighs other alerts from the same machine.
Incidents Investigation
The AI SOC
Kandaro Investigation
Coming soon
Will investigate alerts from
- Your EDR and SIEM
- Identity and cloud alerts
- Deception and Emulation
Emulation will test it Coming soon
- Runs real attacker techniques in your environment, on your schedule.
- Will report for each step whether Investigation caught it, and how fast.
- Re-runs the same step after a fix, so you can show that it works now.
Attacks Investigation
Already have a SOC or an MDR provider? Deception feeds it today through your SIEM, a webhook or the API. See Kandaro and your SOC.
Integrations
Fits the tools you already run.
Kandaro does not replace your SIEM or EDR. It delivers incidents to them and to the people on call, and keeps a read-only door open for your own tooling.
Available now
LiveSigned webhooks
Every incident as JSON to your own endpoint, signed with HMAC-SHA256 and protected against replay. Works with any SIEM, SOAR or ticketing tool that accepts HTTP.
Slack
Incidents in the channel your team already reads, filtered by severity.
Email
Readable incident emails to people or shared inboxes.
SIEM connectors
Incidents delivered to Splunk, Elasticsearch and Microsoft Sentinel as OCSF findings with MITRE ATT&CK techniques.
API and MCP server
Read-only API keys for your own tooling, and an MCP server so an AI SOC agent can ask Kandaro what happened.
JSON export
Export your events and incidents at any time. Your data, your exit.
On the roadmap
Planned- Next
More destinations
Syslog and CEF for other SIEMs, and Microsoft Teams, Jira and PagerDuty for the people on call.
- Then
API for managing decoys
Write access to the API, so you can create and change decoys and alert channels from your own tooling.
- With Emulation
EDR and SIEM read access
Read-only connections that let Emulation check automatically whether your SOC blocked or detected each attack step.
Planned items have no dates yet, and the order may change. Which destination or EDR should come first? Tell us when you join the waitlist.
Security-first engineering
We sell security. We can’t be the breach.
Our sensors run inside your network and our platform holds intelligence about it. So we build Kandaro the way we’d want a vendor to build it.
Sensors take data, never code
Nothing we send can make a sensor in your network execute anything.
MFA is not optional
Passkeys or TOTP for every user. Staff use passkeys only.
Isolation in the database
PostgreSQL row-level security behind every tenant query.
Signed, pinned images
Cosign-signed sensor images with an SBOM. No silent updates.
Attacker data is hostile
Rendered as text only, escaped per channel, fenced off from the AI.
EU-hosted, GDPR
EU infrastructure, local AI by default, 90-day retention.
FAQ
Frequently asked questions.
The six we hear most. The rest are on the FAQ page.
What is Kandaro?
What is deception technology?
How is Deception different from a SIEM or EDR?
Does Kandaro integrate with my SIEM or EDR?
Can Kandaro stop an attack, or only detect it?
Does Kandaro offer a SOC?
Emulation, data, price and running sensors are in the full FAQ.
See your first explained finding today.
Start a 14-day trial of Kandaro Deception. Your first decoy is running in about a minute.
Card required · cancel before day 14 and pay nothing