Skip to content

Product

Everything about your clouds, in one read-only place

AssetHermit reads your AWS, Azure and Google Cloud estates with keyless, read-only access, and turns them into one inventory with its history, its findings, its costs and its operations.

AWSAzureBetaGCPBeta

Inventory and search

Every asset of every cloud in one inventory, searchable in one query language.

Connect an AWS Organization, a GCP organization or an Azure tenant once. Every account in it, and every account added later, lands in one normalized inventory.

  • Search every cloud at once

    One query language for AWS, Azure and GCP. It completes properties and values as you type, or you build the query in a form.

    Read the docs: Search every cloud at once

  • Every property of every type

    Each resource type has a view with every property its cloud returns as a column, so no detail needs the console.

    Read the docs: Every property of every type

  • Asset groups

    Name a set of assets by a query, such as Production or Payments team, and narrow the inventory, findings and costs to it.

    Read the docs: Asset groups

  • Find anything

    Press ⌘K for any asset, account, rule, page or setting by name, across the whole workspace.

  • Your AI estate

    Bedrock, SageMaker, Vertex AI and Azure AI resources as one set of AI types, with 21 rules of their own.

    Read the docs: Your AI estate

  • CSV exports

    Export any search, the change feed or a rule's findings. Large exports build in the background and wait for you.

    Read the docs: CSV exports

Changes and code

What changed, who changed it, and whether it went through your infrastructure code.

Every sync compares each asset with the last, so the inventory is also a history. AssetHermit tells routine noise from real changes, and changes in code from changes by hand.

The Changes page: each change with the properties that changed, who made it and with which tool, and filters for changes made by hand, outside code or during a freeze
  • Full change history

    Every create, change and delete, with the properties that changed. Timestamps and counters that change on their own are folded away.

    Read the docs: Full change history

  • Who changed itBeta

    Each change tied to the calls that made it, from CloudTrail, Azure's resource changes and Google's audit logs: who, with which tool, by hand or not.

    Read the docs: Who changed it

  • Infrastructure as code coverage

    Which assets CloudFormation, CDK or Azure deployment stacks manage, how much of what you expect is covered, drift, and who changed a managed asset by hand.

    Read the docs: Infrastructure as code coverage

  • Terraform stateBeta

    Connect the buckets your Terraform or OpenTofu state lives in. Only the fields that tie a resource to its address are kept.

    Read the docs: Terraform state

  • Change calendarBeta

    Certificate expiries, key rotations, runtime and version end of support, commitments ending and provider maintenance, on one calendar.

    Read the docs: Change calendar

  • Planned changes and freezes

    Put planned work and change freezes on the calendar, see changes that landed in a freeze, and subscribe from your own calendar app.

    On the Team and Enterprise plansRead the docs: Planned changes and freezes

Security findings

175 audit rules after every sync, each with the fix, tuned to what matters to you.

The same query language drives the audit rules, so every finding is one click from the inventory, and you can read exactly what a rule checks.

The Findings page: open findings by severity, a note that most come from a few patterns to tune, and each rule with its findings and accounts
  • 175 audit rules

    Public buckets, open ports, unencrypted disks, old keys and more, checked after every sync. A finding resolves when a sync no longer matches it.

    Read the docs: 175 audit rules

  • A guide with every rule

    Why it matters, when it's fine, and the console steps and command that fix it. Findings one command fixes are grouped together.

    Read the docs: A guide with every rule

  • Is that port really reached?

    Rules on security groups open to the internet show what your flow logs saw: reached, attempted or quiet.

    Read the docs: Is that port really reached?

  • Tuned to your estate

    Environments from your groups, tags or accounts, lower severity in sandboxes, and accepted findings with a reason, an owner and an expiry.

    Read the docs: Tuned to your estate

  • A tuning assistantBeta

    It finds the noise, such as a rule that fires everywhere or a finding that flaps, and proposes changes an admin approves.

    Read the docs: A tuning assistant

  • Evidence for audits

    Export a rule's findings with their exceptions and reasons, and every change with who made it, for your auditors.

Costs and savings

What every account, team and asset costs, what changed, and savings the bill proves.

Cost comes from your clouds' own cost exports, normalized to FOCUS and linked to the assets in your inventory, so a cost has an owner and an asset has a cost.

The Costs page: the last 30 days' daily cost by service, across AWS and Azure, with totals and each cost source's sync status
  • Cost reports

    By account, service, region, tag, group and asset, in your currency, with the month's forecast, last year beside it and a sentence on what moved.

    Read the docs: Cost reports

  • Savings, priced from your bill

    Unused disks, idle databases, oversized instances and old snapshots, each with what it cost you, never a list price.

    Read the docs: Savings, priced from your bill

  • Anomalies, explained

    Every unusual rise or fall in daily spend, with the service, account and usage behind it, and whether usage or price drove it.

    Read the docs: Anomalies, explained

  • Savings the bill confirms

    Each opportunity is followed until it's gone, then the bill is checked. Only what it shows counts as saved.

    Read the docs: Savings the bill confirms

  • BudgetsBeta

    Fixed or worked out from your own spend, by month, quarter or year, with forecasts that warn before a budget runs out.

    Read the docs: Budgets

  • AI spend by modelBeta

    What Bedrock, Vertex AI and Azure OpenAI cost by model, and the accounts that spend on AI without an AI resource.

    Read the docs: AI spend by model

Cost for finance

Cost centers, month close, commitments and unit costs, for the people who book the bill.

Give every cost row an owner, close the month and book it, and know before a commitment ends.

The Allocation tab: coverage, each cost center's direct and shared cost, and where the unallocated cost is
  • Cost centers and allocation

    Ordered rules give each cost row to a team or product, by account, tag, group or service, and split shared cost fairly.

    On the Team and Enterprise plansRead the docs: Cost centers and allocation

  • Month close

    Close a month once its cost stops changing: each cost center's statement and a journal by GL account, frozen. Restatements are booked once.

    On the Team and Enterprise plansRead the docs: Month close

  • Commitments and contracts

    Every savings plan, reservation and committed use discount, its use, its end and what ends with it; spend commitments burned down.

    Read the docs: Commitments and contracts

  • Unit costs

    Cost per order, per customer or as a share of revenue, from numbers your business counts, with a target.

    On the Team and Enterprise plansRead the docs: Unit costs

  • CarbonBeta

    Your providers' own emissions estimates by service, account and cost center, with a yearly CSV for your sustainability report.

    Read the docs: Carbon

  • Resellers and their customers

    Partner discounts shown or left out, and a cost source that imports only some accounts, so each customer sees only its own.

Operations

Uptime, provider incidents, usage and logs, read from the clouds you already run.

What your clouds already measure, in one place and tied to the inventory: no agents to install, nothing to copy out of your accounts.

The Uptime page: monitors, outages and public endpoints watched, probe addresses, and each monitor's last 24 hours, 30-day availability and latency
  • UptimeBeta

    Your clouds' own health checks and canaries in one place, the public endpoints none of them watches, and every outage with what changed around it.

    Read the docs: Uptime

  • Provider health eventsBeta

    What AWS, Azure and Google Cloud say about your own accounts: incidents, scheduled maintenance and actions due, with the resources they name.

    Read the docs: Provider health events

  • Usage and capacity

    Each asset's metrics: idle and oversized resources, unusual usage, capacity running out and service quotas near their limit.

    Read the docs: Usage and capacity

  • Where every asset logs

    Your log groups and where each trail, flow log, load balancer, bucket and function writes, and what isn't logged at all.

    Read the docs: Where every asset logs

  • Log queriesBeta

    Query CloudWatch, S3, Google Cloud and Azure logs where they are, with an estimate before each query and caps your admins set.

    On the Team and Enterprise plansRead the docs: Log queries

  • Cloud news that affects you

    The providers' announcements, retirements and security bulletins, matched against the services and regions you actually run.

    Read the docs: Cloud news that affects you

AI assistant

Ask about your clouds in plain words, in the app, in Slack or in your own AI client.

The assistant reads only what you may see, shows what it read, and is off until an admin turns it on.

  • Ask in plain words

    Turn a question into a query in the search bar, or ask the panel and get an answer from your inventory, costs, findings and logs.

    Read the docs: Ask in plain words

  • In your AI clientsBeta

    Claude, ChatGPT, Cursor and other MCP clients ask about a workspace signed in as you, with an admin's approval.

    Read the docs: In your AI clients

  • In SlackBeta

    Ask in a direct message, or mention the app in a channel, and the answer streams into the thread.

    On the Team and Enterprise plans

  • On your own modelBeta

    Connect Claude on your own Amazon Bedrock, and the assistant answers there, drawing no credits.

Working together

The right people told, a shared wiki for what you know, and a setup guide for what's next.

Alerts that resolve themselves, a summary each week, and a place for runbooks that show live numbers.

  • Notifications

    Alerts when a sync can't read your cloud or an endpoint goes down, notices when findings open or spend jumps, in your inbox, by email and in Slack.

    Read the docs: Notifications

  • A weekly digest

    Each Monday: what changed, findings opened and resolved, cost against the week before, anomalies, savings and provider incidents.

    Read the docs: A weekly digest

  • WikiBeta

    Runbooks and notes beside the assets, groups and rules they're about, with live lists, counts and cost charts in them.

    On the Team and Enterprise plansRead the docs: Wiki

  • A setup guide

    The next steps worth taking in your workspace, worked out from what's connected and what isn't.

  • Roles and sign-in

    Organization and workspace roles, MFA on every password, and sign-in with Google or Microsoft.

    Read the docs: Roles and sign-in

  • Support from the app

    Open a case from where you are, with the context attached, and follow it to the answer.

    Read the docs: Support from the app

Take stock of your cloud

Start a 14-day free trial, no card needed. Connect a first cloud and see everything you run in one place.