What's new in Vantrex

Vantrex is in pre-release and changes quickly. This is the record of what has shipped, newest first, from the first deployment on 28 October 2026.

  1. Vantrex Analytics

    See what happens in your own app: sign-ins, sign-ups, clicks and who is online, for any Next.js or React app with a server part.

    • NewAnalytics projects. Create one from New project, give it your website's address and choose which details of signed-in people to collect. You get a publishable key for the browser and a secret key for your server.
    • NewA new Analytics section in the dashboard: a home with a card per project and four summaries, Logins and Signups pages with success and failure counts, countries, browsers, devices, time taken and why attempts failed, and an Active users page with a live map.
    • NewThe @vantrex/analytics package for npm: a browser part, a server part and a Next.js component, plus a hook for Better Auth. Logins and sign-ups are only accepted from your server with the secret key.
    • NewA public analytics API at api.analytics.vantrex.cloud for sending events and reading statistics, with access tokens that carry a new analytics scope. Events are rate limited per visitor and per project.
    • NewAnalytics projects can be linked to a deploy project to filter by deployment, and rotate their secret key (the old one keeps working for 10 minutes). Settings include the project ID, which details are collected and a CSV or JSON export of your data.
    • InfrastructureAnalytics is a separate service with its own release and its own status component, so it keeps running if deployments are down, and the other way round.

  2. A new look, and production on AWS

    One visual language across the dashboard, sign-in, status page and error pages, and the move to a two-region AWS stack.

    • ImprovedA new interface called Aurora Glass: translucent panels over an animated aurora, applied to the dashboard, sign-in and device pages, the public status page and the pages visitors see when a site is unavailable. The animation pauses when hidden and stills when your system asks for reduced motion.
    • ImprovedError pages served by the edge now say which region answered (for example Ohio or Sydney) and carry the region in a response header, which makes support questions easier to answer.
    • InfrastructureProduction moved to AWS: Ohio (us-east-2) for the control plane, builds and one edge and runner, and Sydney (ap-southeast-2) for a second edge and runner, behind AWS Global Accelerator. Every deployment is available in both regions, so a visitor is never proxied across the ocean to a server.
    • InfrastructureServers are reached only through AWS Session Manager; no node accepts SSH from the internet. Instance metadata is restricted to version 2 with a single hop, and build and run containers cannot reach the private network or the metadata service.
    • NewThe vtx command-line tool is published for macOS (Apple silicon and Intel), Linux and Windows at releases.vantrex.cloud and downloads straight from object storage.
    • ImprovedScheduled maintenance windows now show on the status page and as a banner in the dashboard, and release notices are emailed to account holders before production changes.

  3. Project templates

    Start a project from a template and deploy it in one step.

    • NewChoose a template when creating a project. Vantrex creates a new repository in your GitHub account or organisation from the template, commits the files, and deploys it. You choose the owner, the repository name and whether it is private.
    • SecurityTemplates are read at a pinned commit and only their own folder is copied. Symbolic links, workflow files and anything over 500 files or 20 MB are refused. Your GitHub token is used for the request and is not stored by the deploy service.
    • ImprovedPages and hosts are served from memory once warm, so a slow database read can no longer slow a visitor's request. A first deployment still appears immediately.

  4. Edge cache, status and operations

    A shared response cache at the edge, and the tools we use to run the platform openly.

    • NewEdge response cache for server-rendered routes. Responses are cached only when your application marks them shareable, keyed per deployment, host, path and variation headers, with stale-while-revalidate, conditional revalidation and purge by project, tag or path. Hit and miss are reported in Cache-Status and a new Cache tab in Observability.
    • NewPublic status page with per-component status, incident and maintenance history, and a JSON feed.

  5. Firewall and security hardening

    A web application firewall for your projects, and a round of account protections.

    • NewFirewall with a managed ruleset (sensitive-file probes, SQL injection, cross-site scripting, shell-command patterns and attack tools; each rule can log or block) and custom rules that match on request properties and can log, allow, block or rate-limit. Events and counts appear in the dashboard.
    • SecurityCustom domains are re-verified on a schedule and stop serving if the DNS record is removed. API keys carry scopes (projects, deployments, environment, domains; read or write) and destructive actions ask for confirmation.
    • NewRequest metrics by country and region, with a map.

  6. Accounts, domains and the audit trail

    Everything around signing in and a project's address.

    • NewTwo-factor authentication, passkeys, password reset by email, and sign-in with GitHub.
    • NewAudit trail of account and project changes: sign-ins, password and two-factor changes, token creation, project, deployment and environment-variable changes.
    • NewCustom domains with automatic certificates, issued the first time the domain is visited once ownership is proven.
    • ImprovedRedesigned dashboard home, project settings and Observability, with avatars, a project logo and a glass logo.

  7. Observability

    See what your sites are doing.

    • NewPer-project and account-wide analytics from the edge: requests, data transfer, status codes, top paths, referrers, browsers, devices, bots versus people, countries and latency percentiles. Data is aggregated; see the Privacy Policy for how long each level is kept.

  8. Runtime logs and previews

    Logs from the running app, and a picture of each deployment.

    • NewRuntime logs for server-rendered apps, in the dashboard and with vtx logs --runtime.
    • NewA screenshot of each deployment, taken in an isolated browser container, shown in the dashboard.
    • NewProject switcher and a redesigned new-project flow.
    • InfrastructureThe job queue moved to a self-hosted Redis on the private network.

  9. Live: Git deploys, Next.js and production

    The first production release.

    • NewDeploy from the command line (vtx deploy) or by pushing to GitHub. Every branch and pull request gets its own preview address, and production changes only when you promote.
    • NewRemote builds for static sites, Vite, Astro and Next.js with live logs, cancel, redeploy, promote and roll back, a lockfile-keyed install cache and encrypted environment variables whose values cannot be read back after you save them.
    • NewNext.js apps with server-rendered routes run in resource-limited containers. A new deployment only goes live after the server has started and answered a check.
    • SecurityBuilds and servers run as an unprivileged user on a read-only file system with no extra capabilities, and with CPU, memory and process limits. Pull requests from forks are never built. Platform secrets are never present in a build's environment.
    • InfrastructureProduction serving from Cloudflare-proxied dashboard and API hostnames and a wildcard certificate for deployments, with content stored in Cloudflare R2 and cached in memory at the edge.