We built everythingbut the idea.

Describe it to Claude. It goes live in your own AWS account, with all 54 built in.

54 built in. One left for you.Pick any element to see what it does.

Element 55 starts with one sentence.

Describe the part only your business has. Claude builds it on Langrove, and the other 54 are already there when it lands.

You, to ClaudeBuild a returns tracker where each store manager sees only their own store's returns, and people sign in with Microsoft.

returns.yourcompany.com

ReturnsSam · Store 12

KettleDamagedOpen
Desk lampWrong colourOpen
Office chairArrived lateRefunded

Sam sees Store 12's returns, and no one else's.

Already built in OiOIDC RlRow security DmDomains WfFirewall RvRevert TmTelemetry +48more

Four people get their week back.

The one who built it

I built it in an afternoon. It's been stuck on my laptop for three months.

Today

Sam, OperationsCan the whole team use your returns tracker?

It only runs on my laptop.

With Langrove

returns.yourcompany.com

22 people signed in this week

IT and security

Everyone's building with AI. I can't approve fifty apps I can't see.

Today

FinanceCan we use this AI app?

OpsPlease approve my tracker

SalesIs this one safe?

44 more waiting

With Langrove

Approved once, for every team

Runs in our own AWS account

Every permission it holds, listed

The developer

Every project starts with sign-in, roles and a deploy pipeline. Again.

This sprint

Sign-in, again

Roles, again

Deploy pipeline, again

The actual feature

This sprint, with Langrove

The actual feature

Sign-in

Roles

Deploy pipeline

The one who signs it off

We need AI to make us faster, without hiring a platform team.

Hire for it

A platform engineer£80,000a year, UK median

With Langrove

Every team, every tool£12,000a year

Seven of the 54, in 25 seconds.

Your demo. Seven pieces missing.

Your demo is missing seven pieces.

Your own address, with a certificate.
Company sign-in with Microsoft, Google or an email link.
Everyone sees only their own records.
API keys kept in a vault, out of the code.
Only a person can ship a release.
It runs in your own AWS account, in Europe, and keeps serving without Langrove.
It works inside Claude, as yourself.

Result: live for the whole company.

Every project we built started with the same list.

For over a decade we built it by hand, client after client. Then we built it one last time, properly, and put it under every app you'll make.

Your AWS accountIt runs there, hosted in Europe.
Your codeRead it and change it any time.
Runs without usRemove our access and your apps still serve.

Every app you build makes the next one faster.

One team builds the mail service. Every other app just uses it.

mail.sendoffered by the Integrations app

Offer a service

Mark what an app offers: its code, tables, flows or routes. Your other apps call it by name, like their own code.

Share central data

Several apps read and write one central table, so a customer exists once, not three times.

mail.receivedSupport starts

React to each other

When something happens in one app, flows in the others start.

Within one organisation. Each app stays on the version of a service it took until someone updates it, so nobody's app changes under them.

IT says yes once. Then every tool can ship.

EUROPE

In your AWS account

Every app runs in one AWS account you control, hosted in Europe. Delete our role and they keep serving.

Ship v1Claude: no deploy tool

Only people deploy

Claude and other AI editors have no deploy or rollback tool. You choose who can ship.

<your-stack>-accessPlaceSecretsButNeverReadThemManageCompiledBundlesAttachDomainsToThisDistribution

Everything we can reach, listed

One IAM role. Every permission it holds is written down in our security review.

What IT gets Security review

£12,000 a year. Every person, every app, all 54.One price for your whole organisation. No charge per person, app or request, and your AWS usage stays on your own AWS bill.

See pricing

Bring the idea.

Request early access

For IT

Approve it once. Stay in control.

Your teams are already building tools with AI. Langrove gives those tools one home in your company's AWS account, where you decide who builds and who ships, and can see exactly what we can reach.

What you control

Tools off personal laptops, into your account.

One account for every tool. Every project runs in one AWS account you control, hosted in Europe, on your company's domain.
You decide who builds. An admin approves each person, sets no access, view or edit on each project, and chooses who can create projects.
Removing someone is instant, including the AI editors they connected.
The AI editor can't deploy, and its table tools only see test rows. Shipping takes a person with edit rights.
Everything we can reach is written down. One IAM role, every permission listed in the security review.
Apps keep running without us. Delete our role and your teams' tools keep serving.

Try it: switch Langrove off and watch the requests. Swipe the diagram sideways to see all of it.

Green dots are your users' requests, answered inside your account. The dashed line is our one way in.

Stays in your AWS account

Your tables and every row in them
Files and the document search index
Your users' sessions and passkeys
Logs and metrics, in your CloudWatch
Your users' requests, answered there

Reaches Langrove

Your app's definitions: models, code, rules and flows
Test rows your AI editor works with
What you open in our console, read live through our role
Your team's accounts and daily usage per project

Our role can read your data. We use it to change your tables and show your data in the console, and CloudTrail records its sessions. Agent prompts go to your model provider, search reranking to Frankfurt unless you switch it off, and sign-in emails through us until you add your own sender. Read the full security review.

For your reviewerThe full security review For the builderHow it works Ready?Request early access

Security review

Everything we can reach, written down.

For the person reviewing us: one page you can attach, checked against the stack template and the code on 29 September 2026.

Where it runs
What reaches us
Our role
The AI editor
If we disappear
Who does what
Updates
Not yet

Where each part runs

Your tables (Aurora DSQL), files (S3), search index (S3 Vectors), secrets (Parameter Store), code, agents and schedules run in your AWS account in Europe, across availability zones.
The front door is CloudFront, which is global. Its firewall and certificate are created in your account in us-east-1, so that region must be allowed.
Langrove's console, compiler and AI editor connection run in Langrove's own AWS account.
Encryption at rest uses AWS-managed keys; there's no customer-managed KMS key yet. Traffic uses HTTPS with TLS 1.2 or later.

What stays in your account, and what reaches us

Stays in your AWS account

Your tables and every row in them
Files and the document search index
Your users' sessions and passkeys
Logs and metrics, in your CloudWatch
Schedules, queues and live-update channels
Your users' requests, answered in your account

Reaches Langrove

Your app's definitions: data models, code, access rules and flows
Test rows your AI editor reads or writes
Anything you open in our console, read live through our role
Your team's accounts, and who may edit which project
Daily usage per project: requests and run time
Credentials you enter for sign-in providers and outside APIs, kept encrypted in our database as well as placed in your account
Sign-in emails, for projects without their own email sender
What you type into the console's chat assistant, which we send to Google Gemini

Also leaving your account: agent prompts go to your model provider with your key, and document search sends the question and candidate passages to Amazon Bedrock in Frankfurt to rank results. Each knowledge base can switch that off.

Everything our role can do, from your stack's template

<your-stack>-accesssessions last at most 1 hourneeds an ExternalId

Statement	Actions	Scope
ManageCompiledBundles	s3:PutObject, s3:GetObject, s3:DeleteObject	the releases folder of one bucket
ListBundlesOnly	s3:ListBucket	that same folder
PlaceSecretsButNeverReadThem	ssm:PutParameter, ssm:DeleteParameter	this stack's parameters; no GetParameter
InvokeThisRuntime	lambda:InvokeFunction	the router and the executor
ApplySchema	dsql:DbConnectAdmin	your cluster, as admin
CustomDomainCertsAndAliases	acm:RequestCertificate, acm:DescribeCertificate	any certificate in the account; ACM can't scope this sooner
AttachDomainsToThisDistribution	cloudfront:GetDistribution, cloudfront:GetDistributionConfig, cloudfront:UpdateDistribution	this stack's distribution, its whole configuration
ManageFlowSchedules	scheduler:CreateSchedule, scheduler:UpdateSchedule, scheduler:DeleteSchedule, scheduler:GetSchedule	schedules named devcore-*
ListScheduleNames	scheduler:ListSchedules	names of every schedule in the account
PassOnlyTheSchedulerRole	iam:PassRole	one role, to EventBridge Scheduler only
ReadTelemetry	logs:FilterLogEvents, logs:GetLogEvents, logs:StartQuery, logs:GetQueryResults, logs:DescribeLogStreams, logs:DescribeSubscriptionFilters	five log groups
ReadMetrics	cloudwatch:GetMetricData, cloudwatch:ListMetrics	every metric in the account; AWS can't scope this

It can read your data and logs. We use it to change your tables and to show your data in our console. Nothing notifies you when we do.
It has no permission to read your secret store. It can publish code to your runtime, though, and code that uses secrets receives its project's secrets when it runs. Treat the role as able to reach them.
It can't create roles or change a policy, and it has no CloudFormation access. The one role it can hand to EventBridge Scheduler can only put messages on your turn queue.
CloudTrail records its sessions and management calls. To log its S3 reads and function calls too, add S3 and Lambda data events to a trail of your own.
Delete the role to cut off our access. Your apps keep serving. Sign-in emails we send for you continue, because they use a separate key, until you give each project its own sender.

What the AI editor can and can't do

It connects over MCP as the person who connected it, with that person's view or edit rights on one project. Removing the person cuts it off at once.

It can't

Deploy or roll back. It has no tool for either.
Use the secrets screen, or delete a project or an access
Upload, move or delete your website's files
Read or write live rows through its table tools

It can

Change code and table structure on dev, which shares production's tables. A removed column is kept 15 days.
Create or edit a sign-in or API access, including setting its secret
Run test calls that use your real keys and accesses
Read and end your agents' live conversations. That's a gap: conversations aren't fenced to test data yet.

If Langrove disappears

Keeps working

Every request to your apps
Sign-in, for projects with their own email sender
Agents, flows and schedules
Document search

Stops

The console and the AI editor connection
Deploys and new runtime versions
Sign-in emails we send for you

Already yours

Every table, file and secret
Each release, as a readable file in your bucket
The stack, running in your account

This holds for a runtime in your own AWS account. In an account we run for you, we are its administrator.

Who looks after what

	AWS	Langrove	You
Hardware, network, regions	Runs them
Lambda, Aurora DSQL, S3 and the other services	Runs them
The runtime's roles, limits and defaults		Designs and publishes each version	Applies updates
Change checks, sign-in, access rules, the sandbox		Builds them
Your data model, code and access rules			Decides them, with your AI editor
Who may sign in to your apps		Provides the methods	Configures them, manages users
Backups		None set up by the stack	Sets them up

How updates reach you

We publish runtime versions. Your AWS admin applies each one to both stacks with a CloudFormation update.
The change set shows which resources change, not what new code does, and versions aren't signed.
New database columns are added by our role when we roll out a version. They're additive, so your current code keeps working.
Rolling back a release moves a pointer and rebuilds nothing. Every router picks it up within 30 seconds. It doesn't undo data or table changes.

What we don't have yet

SOC 2 or ISO 27001: No date yet; Every permission we hold is in your stack's template, and CloudTrail records our role's sessions in your account.
Independent penetration test: None published; The stack template and the role above are open for your own review.
Uptime commitment: None; Requests are answered in your account, so an outage on our side stops editing, not serving.
A second region: None; Everything runs in Europe, across availability zones.
Backups set up for you: Not by the stack; Your data is in your account, so set backups there. The file bucket is versioned.
A hard sandbox: Planned; Code runs in Node's vm module, which Node doesn't treat as a security boundary, and all projects in an organisation share one runtime.
Signed updates: None; You apply each one yourself, after reading its change set.
An alert when our role reads data: None; CloudTrail records its sessions; data events on your own trail add S3 reads and function calls.

Still decidingQuestions Ready?Request early access

Pricing

One price. The whole company.

Ten people or two hundred, one tool or forty: the same £12,000. Build as much as you like.

In your AWS account

£12,000a year, per organisation

Unlimited people and projects
No charge per user or per request
Every feature and every update

Request early access

AWS usage and AI model costs are billed to you directly by AWS and your model provider.

In an account we run

£200a month, plus hosting

No AWS account needed
A dedicated account, hosted in Europe
Set up and managed by us

Request early access

Hosting is billed by what your apps use, with a monthly minimum.

Or build it yourself

What a platform engineer would build, already built.

Scalable, secure and ready for AI, from day one.

One platform engineer£80,000 a year

Langrove£12,000 a year

Everything that engineer would set up. All 54 capabilities, included.

Scalability

Serverless hosting that scales by itself
No servers to size or patch
Spread across availability zones
A global front door

Security

Company sign-in, with passkeys
Deny-by-default permissions
Secret store for API keys
Web firewall, rate limits and encryption

AI, built in

AI agent setup
Vector search over your documents
Every app usable from Claude
Your own model keys

Operations

Releases with 30-second rollback
Telemetry: logs, metrics and alerts
Scheduled and background jobs
Live updates on every screen

Fully customisable. Secure by default.

Your data model, your code and your access rules are yours to change, for every app. The guard rails stay on while you do.

Built on AWS services you already trust

CloudFrontAWS WAFCertificate ManagerRoute 53API GatewayLambdaAurora DSQLS3S3 VectorsParameter StoreSQSEventBridge SchedulerAppSync EventsCloudWatchSNSBedrockIAMCloudFormation

The salary is the median advertised for a UK platform engineer in the six months to September 2026 (ITJobsWatch), before employer costs. AWS usage is on your own bill either way. Still yours to do: own the AWS account, apply our updates and set up backups.

Ready?Request early access Still decidingQuestions

Questions

Questions, answered straight.

What exactly is built in?: 54 capabilities across sign-in, security, data, AI, logic, integrations, shipping and hosting. The full list is on the home page, each with one line on what it does.
Can our apps share data and features?: Yes, within one organisation. An app can offer its code, tables and events as a service, and your other apps use them. Each app stays on the version it took until someone updates it.
Can I bring a prototype I've already built?: Your front end, yes, if it builds to static files. The backend is rebuilt on Langrove: describe it to Claude Code and it builds it here, with sign-in and access already in place.
Will it scale?: It runs on AWS Lambda, Aurora DSQL and S3, which grow with use, so there are no servers to size. The defaults, which your AWS admin can raise, are 200 requests a second with bursts of 400, and 10 background jobs at once. A single request gets 15 seconds; longer work runs as a background flow.
Do I need to be a developer?: You need to be comfortable directing Claude Code in plain words and checking what it built. You don't write the code yourself. Someone in IT sets up the AWS account once.
Which AI editors work?: Claude Code is the one we test end to end. The console also gives setup steps for Cursor, VS Code and ChatGPT.
What is the code written in?: JavaScript, run in your account. You can read every line in the console.
Where do prompts go?: What you type in Claude Code goes to your model provider, under your account, and Langrove receives the changes it makes. Your agents send prompts to the provider you choose, with your key. The console's own chat assistant sends what you type there through Langrove to Google Gemini.
Can we leave?: Your data and files are already in your account, and each release is a readable file in your bucket. There's no export to a standard codebase yet, so moving means rebuilding the app's logic.
What will AWS cost?: It mostly grows with use, and shows on your own AWS bill by service. The firewall and DNS zone cost a little even when idle. We haven't published a measured typical figure yet.
What happens if we stop paying?: Your apps keep serving. The console, editing and deploys stop.
Who looks after the runtime?: It's built from managed AWS services, so there are no servers to patch. We publish runtime versions, and your AWS admin applies each one.
We're an agency. How does it work for us?: Build any number of client apps on one price. Apps in one organisation share your AWS account and one runtime; a client who needs their own account needs their own organisation.

A great fit if you

have an AWS account and someone who can administer it
have people who already build with Claude Code
have tools stuck between demo and real use
need company or client data to live in your own account

Not a fit yet if you

need SOC 2, ISO 27001 or an uptime commitment today
must keep data in the UK, or in two regions
review every change as a pull request in git
need web pages rendered on the server

Bring the idea.

Request early access

Privacy

What this site collects, and why.

This page covers langrove.build only. The Langrove console and the apps built on Langrove are separate services. Last updated 1 October 2026.

Who we are

[Company legal name], [registered address], runs this site and decides how its data is used.
Write to [privacy email] about anything on this page.

What happens on every visit

Your cookie answer is remembered. We keep it in your browser's local storage for six months so we don't ask again. It is not a cookie and it is never sent to us.
Nothing loads from anyone else. Fonts and scripts come from this site's own server, not from Google or a CDN.
Our host sees your request. Amazon Web Services, which hosts this site in Europe, receives your IP address and browser details to deliver each page and to block attacks. We keep these request logs only as long as security and troubleshooting need. Our basis is our legitimate interest in running a working, safe site.
"Send this to IT" stays on your device. It copies a note to your clipboard; nothing is sent to us.

Only if you say yes: Microsoft Clarity

Clarity shows us how people use this page, as heatmaps and replays of visits, so we can see what is unclear. It doesn't load until you choose Accept. Right now it is off for you.

What it records: the pages you view, clicks, scrolling and mouse movement, your screen size, browser and device, and a rough location worked out from your IP address. Clarity masks what you type into fields.
Who receives it: Microsoft, which runs Clarity. Microsoft may process it in the United States and is certified under the EU-US Data Privacy Framework. We tell Clarity you have not agreed to advertising storage. Read Microsoft's privacy statement.
How long: Clarity keeps recordings for 30 days and its reports for up to 13 months.
Our basis: your consent. Withdraw it any time with Cookie settings at the foot of the page; Clarity stops and its cookies on this site are deleted.

Cookie	Set by	Kept for	What it does
_clck	this site, for Clarity	1 year	Recognises you as the same visitor next time
_clsk	this site, for Clarity	1 day	Joins the pages you view into one visit
CLID, MUID and others	Microsoft, on its own domains	up to 1 year	Microsoft's own identifiers for Clarity

Your rights

You can ask for a copy of the data we hold about you, and ask us to correct it, delete it, limit how we use it, or stop using it. Write to [privacy email] and we'll answer within one month.
You can also complain to a data protection authority. Ours is [supervisory authority], and you can complain to the one where you live.
When what this site collects changes, we update this page and the date at the top.