drifted

Your Ready for every change.

Ship faster. Spend less time recovering.

Run end-to-end checks before and after code or config changes. Catch failures early, get the evidence to investigate, and verify the fix.

Run your first workflow for free.

Talk about enterprise

A PASSING BUILD IS ONLY THE START

The build passed.
The customer flow didn’t.

Follow the failure to a fix
release-check / priority-orderILLUSTRATED RUN
Add 10 field kits
Choose overnight shipping
Place the order
Verify fulfillment quantity
Saved fulfillment quantity
Expected10 units
Received1 unit

The page said “Order confirmed.”

Failed assertion → source change → verified repair
Browser + API validation CI to production Engineers in controlSee a failed change

WHEN A CHANGE GOES WRONG

Every minute
has a cost.

Lost transactions. Interrupted work. Engineers pulled into an incident. Drifted is built to shorten the time between a bad change and a working system.

TIME TO DETECT

Find the failure early.

Run the workflows customers depend on while the change is still fresh. Catch a broken path in staging or in the checks after a deploy.

TIME TO ROOT CAUSE

Start with the evidence.

Go straight to the failing step, the expected result, and what actually happened. Bring that evidence to the source instead of starting from scratch.

TIME TO RECOVER

Know the fix worked.

Review a proposed code change in a draft PR. After a fix or rollback, run the same workflow again to verify the customer path is working.

EVERY CHANGE NEEDS A CHECK

More diffs.
Same responsibility.

Developers, automation, and agents can all move faster.
Your customers still expect it to work.

A developer’s commit. An agent’s pull request. A config update on a Friday. Each can change how the system behaves. Validate the workflows that matter before and after those changes, so teams can keep shipping with confidence.

Code Agent PRs Configuration

ONE CHANGE, FROM FAILURE TO REPAIR

The order looked right.
Fulfillment got one item.

A config change cut a 10-item order down to one. Follow the failed assertion, inspect the proposed repair, and see the same workflow pass again.

Interactive walkthrough illustrative run
demo app / orderSandbox
Atlas Supply

Equipment for the field.

Response kit

OVERNIGHT
10 units

Rugged field kits for your response team.

Order confirmed
Demo order · no real purchase
What the server actually savedAPI response
fulfillment.quantity10 units As expected
drifted/ Order fulfillmentPassed
HEALTHY BASELINE

An order that works, end to end.

The confirmation page, order record, and fulfillment queue all agree.

  1. 01–03Add 10 field kits · choose overnight shipping
  2. 04Confirm the order page shows 10 units
  3. 05–07Read order · verify status and shipping
  4. 08Verify fulfillment quantity is 10
  5. 09Confirm shipping is overnight
9 of 9 steps passedReady for the next deploy

Start with a journey that proves the result: an order for 10 field kits should send 10 units to fulfillment.

Same order. Correct quantity. All 9 checks pass.

See what Drifted validates

END-TO-END, BEFORE AND AFTER RELEASE

Check what your
business depends on.

A login, an order, a payment, a provisioning flow. Define what success looks like and keep testing it as the system changes.

01

Validate the whole workflow.

Follow a customer action through the browser and your APIs. Check the saved result, not just whether a page loads or an endpoint returns 200.

Browser + APIOutcome assertions
02

Put checks in the release path.

Run workflows from CI before or after a deployment. Use the result in your pipeline to decide whether a rollout should continue.

GitHub ActionsCI pass / fail
03

Check across environments.

Keep staging and production workflows separate. Run public checks in managed browsers and use private runners for internal services.

Environment controlsPrivate runners
04

Keep checking after deployment.

Schedule repeat runs, keep an incident history, and send failures and recoveries to an HTTPS webhook. A release check is the beginning.

Scheduled validationWebhook alerts
Security checks for headers, exposed files, and keys Incident history and webhook alerts

THE ENTERPRISE ROADMAP

From the pull request
to the on-call team.

We’re building toward a complete validation loop: connect a change to the failed workflow, route the incident to its owner, and verify recovery. Help shape it for your team.

AVAILABLE

GitHub Actions

Run workflow checks in CI and use the result in your release pipeline.

AVAILABLE

GitHub

Connect repository source, inspect proposed fixes, and open draft pull requests.

AVAILABLE

Alert webhooks

Send signed failure and recovery events to your own HTTPS endpoint.

PLANNED

PagerDuty

Native incident routing to bring validation failures into your on-call process.

Bring us the workflow you can’t afford to break.

Let’s talk about your environments, release process, and incident response.

Talk about enterprise

Your workflows.
Your control.

Our security vision: validate critical journeys with less data, limited access, and clear accountability.

Talk security with us

Available today

  • Encrypted credentials

    AES-256-GCM encryption, with test access scoped to your environments.

  • Isolated execution

    Sandboxed browser runs and configured secrets redacted from saved results.

  • Optional AI

    Checks run without AI. Requested AI assistance shares relevant context with a provider.

On our roadmap

  • More team control

    SSO, roles, access audit logs, and configurable data retention.

  • Secrets stay with you

    Customer-controlled keys and private runners using your own secret manager.

  • Independent assurance

    Penetration testing and a SOC 2 audit are on our roadmap.

How we handle your data

PRICING

Pay for proof.
Not for seats.

Every plan meters verification runs: the checks your agents, pull requests, and schedules trigger. Invite the whole team. No per-seat fees.

2 months free with annual billing.

Free

$0 / forever

Install the agent hooks. Prove one journey.

  • 1 app, 3 workflows
  • Daily scheduled checks
  • 200 verification runs a month
  • CLI, MCP server, and GitHub Action
  • Evidence API for Claude Code and Codex
No credit card needed.
LAUNCH PRICING

Pro

$29 / month

Launch price. Pro moves to $79 a month when Team ships; subscriptions started now keep their rate.

  • 5 apps, 25 workflows
  • Hourly scheduled checks
  • 2,000 verification runs a month
  • Preview URL targets for pull requests
  • Performance verdicts against your baseline
  • Reviewed repair proposals
$29 billed monthly. Or pay $290/year and save $58.
Manage or cancel in your account.
EARLY ACCESS

Team

$399 / month

Every pull request proves it is done before it merges.

  • Unlimited apps, 200 workflows
  • Checks every 15 minutes
  • 10,000 verification runs a month
  • Required status checks on pull requests
  • Repair pull requests opened for you
  • Performance regression gates
Request early access Early access opens this quarter. Everything in Pro is included.
[email protected]

Enterprise

Custom

Annual coverage contracts for the journeys your business runs on.

  • Priced per covered journey
  • Journeys maintained by Drifted
  • Private runners and preview environments
  • SSO, audit log, and retention controls
  • Incident routing and SLA
Talk about enterprise For teams covering 50 or more journeys.
[email protected]

All plans include browser and HTTP checks, run evidence, the CLI, the MCP server, and the GitHub Action. A verification run is one check triggered by an agent, a pull request, a schedule, or a person. Additional runs cost $0.03 per browser run and $0.003 per HTTP run; overage billing begins when Team ships. Limits apply across your account; the included homepage workflow counts toward your total.

BEFORE YOU CONNECT

How it fits
your team.

Where does Drifted fit in our release process?

Run the same critical workflows in staging, from CI, and against production. The CI command returns a pass or failure your pipeline can act on. Scheduled runs keep checking between releases. Your team decides which checks gate a rollout.

Does this replace our tests or observability?

It works alongside them. Unit tests check individual pieces; logs and metrics help explain what a system is doing. Drifted checks whether a complete customer workflow still produces the right result, and gives engineers the failed step to investigate.

What about config changes and AI-generated code?

A workflow check validates behavior regardless of who wrote the change or whether it changed code or configuration. Trigger checks from your deployment pipeline and keep them running afterward. Automatic attribution to a specific code or config change is part of the enterprise roadmap.

Can we connect PagerDuty?

Native PagerDuty integration is planned. Today, Drifted sends signed failure and recovery events to a configured HTTPS webhook. Talk to us about the incident routing your team needs.

Will Drifted merge a fix or roll back production?

No. Drifted can prepare a source-based repair and open a draft pull request on a separate branch. Your team reviews the change and controls merges, deployments, and rollbacks. Run the workflow again to check the result.

THE NEXT CHANGE IS ALREADY ON ITS WAY

Give your team
the confidence to ship.

Start with a critical workflow. Know it works after the next change.

Talk to us