Cloud and AI security that finds, proves and fixes.
Posture for AWS, Azure and Google Cloud, with attack paths, code and secrets, AI agents, real-time threats and fixes as pull requests.
No signup. Read only. Results in minutes.
SOC 2 and ISO 27001: in progress. SOC 2 letter · ISO 27001 letter
Find
Attack paths to your customer data
Cloud and AI agent paths, each hop checked in your account.
- Public endpointOpen to the internet
- Your APIRuns as a cloud role
- Cloud roleAllows every action
- DatabaseCustomer data, no backup
AI agents in your cloudFound with what each can reach
LiveWhat your agents doTheir calls and the data they touch
Live
Prove
Hijack your own AI agent, safely
Prove it plants one line in a file your agent reads, then reports if it obeyed.
- Agent read a file with one planted line
- Agent wrote the file the line asked for
- Critical finding raised
Block
Stop the delete that cannot be undone
The gate can block a destructive call before it runs.
- Outside contentA storage bucket another account can write
- Your AI agentReads it, holds a delete tool
- Agent gateopt-inCan block the delete before it runs
- DatabaseCustomer data, cannot be undone
Fix
The fix arrives as a pull request you review
or a review ticket where a pull request is not possible
In your infrastructure code. Fixed in the code that creates it, so a redeploy does not undo it.
- Netlumi opens it
- You review
- You merge
- A redeploy keeps it
data "aws_iam_policy_document" "app_runtime_debug" {- statement {- sid = "TemporaryDebugAccess"- actions = ["*"]- resources = ["*"]- } }
Watch
Leaked keys and live threats
Decoy keys where real ones live. Real-time threats.
- Decoy key usedAlert names the IP it came from24.9 s
- Real-time threatsWithin minutesLive
Also: posture for AWS, Azure and Google Cloud · IaC, secrets and Dockerfiles · SOC 2 and ISO 27001 mapping · audit trail
Questions buyers ask.
What does Netlumi do?
Cloud security (CSPM). AI agent security. Attack paths across cloud and AI. Fixes as pull requests. Blocks for risky agent actions. We scan your AWS, Azure and Google Cloud accounts, your repositories and the AI agents you run, and open the fix as a pull request in your IaC code for your team to review and merge.
Why not just use AWS Security Hub?
Keep Security Hub on; the two work together. Security Hub collects findings about your AWS accounts, and its automated response changes the live resource. Netlumi fixes the IaC code that created the resource, with a pull request you merge, so the fix is still there after the next deploy. Netlumi shows which findings chain into a path to your data and marks an issue critical or high only when the evidence shows it matters in your account, so you get a short list. Netlumi also shows what your Amazon Bedrock agents can reach and change, including actions that cannot be undone, can stop an agent action that would destroy data you cannot restore (the optional gate, Business and Enterprise plans), and uses decoys to catch misuse. And Netlumi covers Azure and Google Cloud, and your code and repositories.
What permissions does Netlumi need?
Read-only. On AWS you create a role with the AWS-managed SecurityAudit policy plus Netlumi policies that grant exactly the read actions Netlumi calls, and Deny policies that stop it reading your data (S3 objects, DynamoDB items, log events, queue messages, secrets). Netlumi assumes the role with an external ID for credentials that expire after one hour and stores no access keys. Azure uses a federated credential and Google Cloud uses workload identity federation, with no stored secrets. For GitHub, a read-only app scans your repositories, and a separate fix app opens pull requests on the repositories you choose.
What data does Netlumi store?
Findings, inventory and attack paths from your cloud accounts; for connected repositories, a working copy on encrypted disks while the repository is connected, and proposed fixes for up to 30 days. Netlumi's search index does not keep the text of your code. Secrets Netlumi finds are recorded by name and a short fingerprint, never in full. Everything is encrypted in transit and at rest, with a key per tenant. Service logs are kept for 30 days.
More questions
Is my data used to train AI models?
No. Netlumi does not use your data to train AI models. Where Netlumi sends code context or findings to an AI provider to write a fix or answer a Lens question, secrets are masked first, and the providers are listed on the Security & Privacy page in the app.
Does Netlumi change my cloud or merge code?
No. Netlumi has no write access to your cloud unless you turn on Apply for me. Fixes arrive as a pull request on a netlumi/fix-* branch; Netlumi never pushes to your default branch and never merges. Your team reviews and merges, and your pipeline applies the change as it would any other.
How many cloud accounts does each plan cover? Does the Starter plan cover 20 accounts?
No. Starter ($249 a month) covers up to 3 cloud accounts, Growth up to 10, Business up to 25, Enterprise up to 50, and the free Developer plan 1. A cloud account is one AWS account, Azure subscription or GCP project. 20 or more accounts means the Business plan; for more than 50, talk to us. Users are unlimited on every plan.
What does "from" mean on Enterprise?
Enterprise starts at $30,000 a year for up to 50 cloud accounts, unlimited repositories and 250 AI agents, and you can buy that size by card today. A larger estate gets a quote, priced by cloud account.
How does the price compare to buying separate tools?
One product and one price. No per-seat charges, no per-resource metering. Teams otherwise buy a cloud posture tool, an AI security tool and a fix workflow separately, and pay for each. Netlumi is one subscription: €0 / €229 / €499 / €999 / €2,300 a month for the Developer, Starter, Growth, Business, Enterprise plans (Enterprise billed annually), with unlimited users. A team with 20 or more cloud accounts is the Business plan, €999 a month for up to 25 accounts, so nobody has to guess what the Starter price covers.
Does Netlumi cover Amazon Bedrock agents?
Yes. Netlumi lists the Bedrock agents in your AWS accounts and shows the data each agent could read, change or delete, including data in another account, and which of those actions cannot be undone. It shows the path from untrusted input, such as a document with a prompt injection, to your data, and the change that breaks it. Agent reach and AI attack paths cover AWS today.
Can I test this on my own agent?
Yes, on the Business and Enterprise plans: Prove it. Plant one harmless instruction wherever your agent takes input (a chat message, a document, a knowledge-base entry, an email or ticket, a web page, a tool result), run your agent as you normally do, and Netlumi shows the write it made, whether it can be undone, and the fix. Netlumi never runs your agent and never reads your prompts.
Which clouds and AI platforms does Netlumi support?
AWS, Azure and Google Cloud for cloud security. For AI: Amazon Bedrock and SageMaker models, endpoints and agents are checked in your AWS accounts; Azure OpenAI and Google Vertex AI usage is detected in your code. The optional gate runs in your AWS account.
What if I have more than one issue at once?
Urgent alerts are rare and reach you at once. Everything else goes into a daily digest. Our automatic remediation workflow keeps fixing issues in the meantime. Each fix arrives as a pull request you review.
Does Netlumi have a SOC 2 report or an ISO 27001 certificate?
Not yet. SOC 2 and ISO 27001 are in progress. Netlumi runs an information security management system aligned to ISO/IEC 27001:2022 Annex A, can walk you through it and answers security questionnaires on request. The letters of engagement for both are on the trust page (netlumi.com/trust#audits). We will publish the reports when they are issued.
What happens when I cancel or leave?
Plans are billed monthly or yearly; a monthly plan has no annual commitment. When a trial ends you pick a plan or drop to the free plan, and nothing is deleted. When you remove the role from a cloud account, its data is kept for 30 days so a reconnect restores it, then purged. If another organisation takes the account over, your copy is purged at once. On request, Netlumi deletes your whole organisation and destroys its tenant key first.
Free to start. Paid plans from €229 a month.
One product and one price. No per-seat charges, no per-resource metering.
What Netlumi will not do
- Change your cloud unless you ask it to.
- Merge a pull request.
SOC 2 and ISO 27001: in progress.
No signup. Read only. Results in minutes.