What we shipped, and when.

The agent is now called pierrr-agent

Sessions that really close, and tabs that answer the click

Automatic certificate renewal now restarts with the server

A profile picture for your organization

Creating a WordPress site moves to the Projects page

A project's backups now contain the database of every WordPress site it hosts

The agent knows how to install a WordPress site

Expired backup archives are deleted again

The agent now cleans up its Docker images after every successful build

Create a WordPress site from a name and an email address

A deployment refused during a platform restart is now replayed automatically, as an…

A build interrupted by a platform restart is now replayed automatically instead of being…

The agent refuses work that reaches it while it is shutting down, instead of accepting it…

Renaming a container now removes the old one

Renaming a project now stops its old containers

A project can now be renamed, not just its label

Volumes now follow a rename

Cleanup of abandoned networks can now detach the entry nginx

Resource inventory now reports the networks tied to the projects database, along with the…

Deleting a project now also takes down the network that linked it to the shared database

Renaming an organisation no longer leaves Docker networks behind

Renaming an organisation is supported

Deploy a specific commit, not just the latest

Fast deploy puts your latest push ahead of the previous one

Deploy the commit of your choice, not just the latest one

An nginx configuration error finally says what is wrong

Security scans can now run on this node

The agent now announces what its machine can take

Public forms are protected against bots, with no friction for you

Your containers' secrets are no longer written in clear in the compose file

Expired archives are really removed from disk

An encrypted backup really is encrypted, or it fails

A build context can no longer escape the repository

Fewer needless reloads of the project nginx

A custom domain could stay on a 404 while everything looked fine on the Pierrr side

Custom domains are served again, and the domain page lists every expected DNS record

Dependency update

Change a container's deploy branch without recreating it

The 14-day trial now collects a card and converts into a subscription

A project now gets its network and its nginx as soon as it is created, without waiting…

A first deployment could sit on a 502 for up to 30 minutes, with nothing visible…

Fixed apps unreachable after their first deploy, network resync action

A restart of the shared agent no longer cancels your builds awaiting an answer with…

Your applications can send mail without holding an SMTP password

The advice given when a container has no /health route is more complete

— each project gets its own network to reach the managed database

A platform update no longer interrupts your builds in progress

The deployment log no longer announces health it has not measured

A build interrupted by a platform update says so clearly

A momentarily unavailable registry on a side service's image no longer fails an…

A deployment now starts the side services its project declares, before the application

When nothing listens on the port, the agent attaches the container's last lines

A container that never turns "healthy" now explains why

New: the TCP ports a project publishes are automatically opened in the host firewall

Per-project network isolation, connecting managed resources, multi-line secrets

Postgres extensions, worker containers, object storage and TCP ports

Fix: scale-app could stay stuck indefinitely on a container name conflict, never…

Critical fix (found in production testing right after 0.20.4-beta): resync-project…

Reliability fix: scaling an instance could "succeed" on the Hub side without the…

Pierrr Security: automated vulnerability scanning

Fix: the agent container image could no longer be built or deployed

MCP: connect several organizations from one client

Fix: a blue/green deployment never came back down to the right instance count if the…

Critical fix (found in production testing, Wheelrn migration): a blue/green deployment…

Health history, drop the period picker

Per-container healthcheck history

Import your data into your managed Postgres

Turn off auto-deploy per container

Two-factor authentication and passkeys

New volume-usage command: the agent can now measure the space a Docker volume actually…

Critical fix (found in production testing): restoring a backup containing a volume always…

Fix: pg_dump's error output was never captured, making any diagnosis of a failed backup…

Critical fix (found in production testing): the scale-app command had never worked, since…

Fix (found in production testing): a command queued behind an ongoing project operation…

New scale-app command: scaling a container to N instances

Scale containers up or down with a stepper, right from the project card

: backups change granularity, from container to project

Crash-loop detection fix: zero restarts wrongly transmitted

Crash-loop container detection

Automatic database env vars, a redesigned topology view, and whole-project backups

Network groups between a project's apps

Network groups for project topology, plus a GitHub reconnect fix

Log retrieval correlated per request

Security dependency updates

Client builds capped on resources

Monitoring CPU consumption fix

Hardening of project decommissioning

Real-time tracking of the agent update

Fix: project decommissioning and resource inventory were ignored

Managed resource inventory

Full decommissioning of a project

Live server update progress and full cleanup on project deletion

502 fix on hosted containers' subdomains

Drive Pierrr with AI: remote MCP server and in-app assistant

Smarter pricing, bulk env editing and in-app communications

Backup restore and download fix

Database backup fix

More reliable backup download

Managed database backups

Backup download fix

Managed backups: restore and download

Zero-downtime deployment (blue/green)

Move to the beta channel

Reliability of the nginx reload on deployment: right after the project's reverse proxy…

(hub↔agent) — One reverse proxy per project instead of one per container

Custom domain to a self-hosted server: the agent now reports your server's public IP…

Automatic edge setup (self-hosted): on one of your servers, the agent can now stand…

Custom (BYO) domains served by the edge: the agent can now expose a domain you own (e.g

Republication: a technical version (no code change) succeeding 0.2.5-alpha, whose…

Republication: a technical version with no code change, succeeding 0.2.4-alpha (whose…

Republication: a technical version with no functional change from 0.2.3-alpha, published…

Internal cleanup: the systemd unit name (pierrr-agent.service) is now centralised in a…

Self-hosted build repaired on hardened hosts: docker build/push failed with "mkdir…

The environment variable log is always there: the "Injected N env var(s)" line now…

Pierrr closed beta is live

The load and start stages are finally distinct: pulling the image (load) and…

Per-container health tracking, based on the real health check: on every heartbeat the…

Security dependencies up to date: golang.org/x/crypto (0.41 → 0.45) and cloudflare/circl…

One compose file per project: the agent maintains a single docker-compose.yml per project…

A clear error when the nginx config is missing: on a deploy-app, if the hub did not…