One command center for everything you run.
Guards inside Fynd. Connectors for the rest. Eight AI agents that detect, investigate and act. A control ladder you set. Everything on this page is the real product running on sample data, so go ahead and press things.
- Security data lake, stored in India
- Eight AI agents that detect, investigate and act
- Your autonomy policy and Fynd's hard limits
- One timeline, one audit log
Eight guards for the paths attackers use.
Each guard runs inside the Fynd platform. Start in shadow mode to see what it would block, then switch to enforce. Try the switches.
- Bot and abuse guard
Scraping, inventory hoarding, fake sign-ups, card testing and coupon abuse.
How: Behaviour scoring at the Fynd edge, device fingerprints, invisible challenges and abuse rules for your brand.
Blocking attacks now - Account takeover guard
Credential stuffing, password spraying and session hijacking.
How: Breached-password checks, login velocity limits, risk-based OTP step-up and session binding.
Blocking attacks now - Checkout integrity guard
Payment skimmers and unapproved or changed scripts on payment pages.
How: Script inventory and approval, tamper detection and a managed content security policy.
Blocking attacks now - API shield
API abuse, token theft and data scraping through APIs.
How: Schema enforcement, per-token rate limits and anomaly detection on Fynd APIs and your extensions.
Blocking attacks now - Admin and insider guard
Compromised or rogue admins, mass exports, price and coupon tampering.
How: SSO and SCIM, just-in-time admin access, IP allowlists, alerts on risky bulk actions and two-person approval.
Blocking attacks now - Data vault
Leaks of shopper personal data.
How: Field-level encryption, tokenisation, bring-your-own-key, masking for support teams, DPDP consent and erasure logs.
Blocking attacks now - AI agent guard
Prompt injection, data leaks and tool misuse in Kaily and your own AI agents.
How: Input and output filtering, tool-call policies and a separate identity and scope for each agent.
Watching only. Logs what it would block. - Store device guard
Tampered POS terminals, rogue apps and stolen store devices.
How: Device health checks, app allowlisting and remote lock and wipe through Store OS.
Blocking attacks now
- DDoS and edge protectionAlways on for every Fynd storefront and API
- Managed TLS certificatesIssued and renewed automatically
- Standard WAF rulesOWASP Top 10 protections at the Fynd edge
- MFA for admin usersRequired for every Fynd console login
- Monthly security scoreReport sent to owners on the 1st
Bring every cloud, data center and app.
Fynd properties are native. Everything else connects read-only first, so you see value before you grant any power.
Native. Nothing to install.
Read-only roles. Response permissions only if you opt in.
Fynd Collector, a small virtual appliance that forwards logs securely.
OAuth apps and API connectors.
Your existing EDR, or one resold through Fynd.
Store OS devices, POS terminals and store networks.
- 1Native on Fynd
Storefront, app, checkout, OMS, Store OS, Kaily, Boltic and the Fynd console are already connected.
- 2Read-only connectors
Clouds, SaaS and identity connect with read-only roles or OAuth apps. Most take minutes.
- 3Fynd Collector
A small virtual appliance in your data center forwards logs from ERP, servers, firewalls and Active Directory.
- 4Response when ready
Grant response permissions per system, such as rotating keys or isolating a laptop through your EDR.
Eight agents. Each with one job.
They work as a team: one removes noise, one builds the story, one acts, one hunts, one drafts the reports. They show their reasoning every time.
Reads every alert, removes noise and groups related signals into one incident.
Builds the full story: who, what, how and how far it spread.
Runs playbooks within your autonomy policy and asks you when it must.
Watches shopper behaviour on storefront, app and checkout in real time.
Finds weaknesses before attackers do: misconfigurations, open services, risky access.
Searches for hidden attackers using fresh intelligence from the Fynd Threat Network.
Keeps evidence, tracks reporting clocks and drafts regulator reports for a human to submit.
Safely attacks your own estate, including your AI agents, every week.
You decide what agents can do alone.
Every kind of action sits on a ladder. Fynd sets the outer limits. You can make agents more careful at any time, and start everything in shadow mode.
Safe, reversible actions agents take on their own in seconds.
- Block attack traffic at the Fynd edge
- Add a challenge or OTP step-up
- Rate-limit an API token
- Roll back a changed checkout script
Agents act, tell you right away, and you can undo in one tap.
- Lock a compromised shopper account
- Rotate a leaked API key
- End an employee's sessions
- Isolate one laptop through EDR
Agents prepare the action and show the blast radius. You approve. Critical ones need two people.
- Disable an admin user
- Isolate a production server
- Change cloud access settings
- Put a storefront in queue mode
Only humans do these. No setting can change that.
- Delete data
- Submit a regulator report
See the blast radius before you approve.
Level 3 actions arrive with what will change, which systems and shoppers it touches, and how to undo it. Critical actions need two people. This card is live: approve it here and the demo updates too.
- Approve in the app, on WhatsApp or in Slack
- Two approvers for critical actions
- Every decision is logged with who, when and why
Suspend admin ravi.k@vastra.co and rotate 6 admin API keys
A login from the attacking proxy pool reached this admin account two days ago. Suspending it and rotating its keys removes any access the attacker may still have.
- Fynd console: 1 admin user
- ERP sync: pauses about 4 min
- Marketing feed: pauses about 4 min
Two integrations use these keys. They pause for about 4 minutes while new keys are issued.
- 1Suspend ravi.k@vastra.co in the Fynd console and Okta
- 2End all active sessions for this user
- 3Issue 6 new admin API keys
- 4Update ERP sync and marketing feed with the new keys
- 5Revoke the 6 old keys
- 6Confirm both integrations are healthy
Ask anything, in plain words.
Ask Fynd Secure knows your incidents, guards, systems and rules. It explains an attack for your CEO, tells you what needs approval and what to switch on next. It answers from your own data and says so when it does not know.
Evidence ready when the clock starts.
The Compliance agent starts the CERT-In and DPDP clocks, drafts the reports and collects evidence for audits. Your team reviews and submits.
6-hour reporting, 180-day logs in India, NTP sync
Rule 6 safeguards and Rule 7 breach process. Obligations start 13 May 2027
6.4.3 script inventory and 11.6.1 tamper detection on payment pages
Evidence collected automatically for 79 of 93 controls
Evidence for 51 of 65 criteria. Audit window opens Jan 2027
Clocks use the sample incident in the demo. This is a summary, not legal advice.
Power, with brakes built in.
Security tools hold a lot of power. These are the rules Fynd Secure runs by.
Security data for Indian brands is stored in India (Mumbai), with logs kept for 180 days to meet CERT-In.
Connections begin read-only. Agents only get response permissions you choose to grant, one system at a time.
Automatic actions stay inside blast-radius limits you set, such as how many shoppers they may affect in an hour.
Deleting data and submitting regulator reports are human-only. No setting can change that.
Each action has the agent's reasoning, the evidence it used and who approved it, in one audit log you can export.
Your data is used to protect you. It is not used to train shared models, and access by Fynd staff is logged.
Be one of the first brands on Fynd Secure.
Early access is open to Fynd customers now. We start in read-only shadow mode, so nothing changes until you say so.