Helpdesk

self-hosted ticket system for your company. Employees raise issues, managers assign them

Language
TypeScript
Stars
0
Last updated

From the README

A self-hosted ticket system for your company. Employees raise issues, managers assign them, officers fix them, and the person who raised the issue confirms the fix. Every person has a role, and admins decide exactly what each role can do. It runs on one Ubuntu server and installs with one command.

Built on Peppermint, customized by Abdul Manan.

Contents

What it does

Built-in roles, from most to least senior (admins can rename them, change what each can do, and add their own; see Roles & permissions):

Role What they can do out of the box
Admin Everything, including roles, permissions and settings.
Senior Manager Sees and assigns everything in their reporting line, organises it, closes issues, sees team reports, and can promote people to roles below their own.
Manager Runs a team: assigns issues to their deputies and officers, closes issues, organises their team and sees team reports.
Deputy Manager Supports a manager: assigns issues to the officers who report to them and sees their reports.
Officer Works on issues: takes them from the queue, updates them, marks them done and sets their availability.
User Raises issues, sees only their own, and confirms a fix or reopens the issue.

How an issue moves: a user raises it → it goes straight to the staff member who handles that user (or waits as unassigned until someone takes or assigns it) → they mark it done → the user confirms (closed) or reopens it.

Teams follow reporting lines. Everyone can report to someone more senior (Officer → Deputy Manager → Manager → Senior Manager). A person's team is everyone below them, so a manager automatically sees and assigns their deputies' and officers' work.

Also included:

  • Deadlines (SLA): every priority has a target time, so each issue gets a due date and late issues are flagged.
  • Your own fields: admins control issue types, priorities, categories and which fields are required (Admin → Issue settings).
  • Analytics: issues raised vs finished per day; open issues by stage; breakdowns by priority, type and category; top clients; workload per person; and a heat map of the busiest weekdays and hours.
  • Reports: by month or any date range, for the whole company, a team or one person. Print to PDF or download as CSV. Officers get their own report.
  • Performance: completion times, done and confirmed counts, reopened issues.
  • Your brand: your logo and company name across the app and the sign-in page (Admin → Branding).
  • Single sign-on: staff sign in with their company account (Microsoft Entra ID, Google Workspace, Okta, Keycloak, ADFS, OneLogin and others) through OpenID Connect, SAML 2.0 or OAuth 2.0. See Single sign-on.
  • On every issue: attachments (drag and drop, image preview, up to 25 MB each), time logged, linked issues (related, duplicate of, blocks) and comments their author can edit.
  • Board and list views: drag cards between columns to change status, priority or type; filters for status, priority, type and person; quick actions from the ⋯ menu.
  • Email: Microsoft 365 (without passwords), Google or any SMTP server for outgoing mail. Email queues turn messages sent to a mailbox into issues, and replies into comments. Everyone chooses which emails they get. See Email.
  • Webhooks: signed messages to Slack, Microsoft Teams, Discord or your own systems when issues are created, assigned, commented on or change status.
  • Documents: runbooks, guides and notes, kept private or shared with chosen roles, named colleagues or the whole company, so work carries on when the person who usually does it is away. See Documents.
  • Equipment: laptops, phones and accessories with your own categories and icons — hand them out, let people confirm what arrived, take requests and read the totals. See Equipment.
  • Customers: a portal where outside clients raise and follow their own issues, and a client list with notes.
  • Notifications: in the app (with a history page) and by email.
  • Forgot password: a 6-digit code by email that works for 30 minutes.

Roles & permissions

Open Admin → Roles & permissions. Everyone has exactly one role; the role decides what they can see and do. The server checks every action, so hiding a button is never the only protection.

  • Edit roles: pick a role on the left and tick what it can do. There are 42 permissions in six groups: Issues, People & teams, Documents, Equipment, Reports & analytics and Administration. Each one says what it allows, and the powerful ones are marked. Turning one on also turns on anything it needs (for example Take issues needs Work on issues).
  • Compare all roles: every role side by side in one grid. Tick boxes across several roles and save them together.
  • New role / Duplicate: create your own roles (for example Team Lead or HR Officer), or copy an existing one and adjust it.
  • Seniority (1–99): sets the order. People can only manage roles, and people, below their own seniority, and can only hand out permissions they have themselves. The Admin role is always at the top with every permission.
  • Default for new accounts: the role given to people who sign up, or sign in with single sign-on, for the first time (the built-in User role unless you change it).
  • Delete: custom roles can be deleted; you choose which role their members move to. Built-in roles can be renamed and changed but not deleted.

Admin → People lists everyone with a search and a filter per role. Change one person's role in place, or tick many people and change all their roles at once. Team shows the reporting lines as a tree and who handles each user's issues.

Every change to roles, permissions, accounts, reporting lines, sign-in settings, webhooks and clients is written to Admin → Audit log with who made it and when.

Documents

Documents is the company's written knowledge: how a printer is reset, which supplier to ring, the steps for a new starter. Write it once and the work carries on when the person who usually does it is on leave.

Every document has an audience, set by its author from the Share button:

Audience Who can open it
Only me Nobody else, except the people whose role lets them review their team's or the company's documents.
Chosen roles and people Everyone with a ticked role (for example every Officer, including people given that role later), plus anyone ticked by name.
Everyone in the company All staff.

Customers on the portal never see documents, whatever the audience.

The list has two tabs: My documents and Shared with me (which reads Shared with me & my team for anyone who can see their team's). Search finds a document by title. Only the author can change or delete a document; everyone else opens it read-only, with the author's name at the top. Opening a document to a wider audience, or deleting someone else's, is written to the audit log.

Five permissions decide who does what (Admin → Roles & permissions → Documents):

Permission What it allows On by default for
Write documents Write their own documents. Everyone, down to User.
Share documents Choose roles, people or the whole company as the audience. Picking people by name also needs See the staff directory; without it they share with whole roles instead. Officer and above.
See their team's documents Open anything written by someone in their reporting line, including private ones. Deputy Manager and above.
See every document Open every document in the company. Senior Manager and Admin.
Edit or delete any document Change or remove other people's documents. Admin.

So out of the box: senior staff publish guidance that officers and users can read, officers write their own runbooks and share them, users keep private notes, and a manager can still open what their own team wrote.

Equipment

Equipment tracks what the company owns and who has it: laptops, phones, keyboards, chairs — whatever you hand out. It answers the everyday questions on its own: who has the spare MacBook, what is free to give a new starter, who still hasn't confirmed the headset that was posted to them, and which warranties run out next quarter.

Categories are yours to build (Equipment → Categories). Each one gets a name, an icon from a picker of 31 (laptop, keyboard, headset, printer, vehicle and so on), a colour, and optionally the person who looks after that kind of kit. Sub-categories go one level deep — Accessories with Keyboards, Mice, Headsets, Docks & cables, Chargers under it. A starter set is already there; rename, re-icon or delete any of it. A category holding equipment can't be deleted until its items are moved.

How an item moves:

  1. Someone with Add and edit equipment adds it. It gets a tag automatically (LAP-0001) unless you type your own, and starts in stock.
  2. A manager hands it to a person. It becomes waiting to be confirmed, and that person is notified.
  3. The person presses I have received it. Now it is with a person, and the date is kept.
  4. When it comes back it returns to in stock; it can also be marked in repair, retired or lost.

Every step is written to the item's own history, so you can read its whole life at a glance.

Asking for equipment (Equipment → Requests). Anyone can say what they need and why. Whoever looks after their team answers: turn it down with a note, or approve it — and pick a free item in the same dialog, which hands it over on the spot. Confirming receipt closes the request off.

Reports (Equipment → Reports) show totals by status and category, who holds the most, items still waiting to be confirmed and warranties ending within 90 days — all limited to what that person is allowed to see.

Seven permissions decide who does what (Admin → Roles & permissions → Equipment):

Permission What it allows On by default for
Ask for equipment Ask for something, and see what they hold. Everyone, down to User.
See their team's equipment What everyone in their reporting line holds, plus what is free. Officer and above.
See all equipment Every item in the company. Senior Manager and Admin.
Hand equipment out Give items to their own reporting line, take them back, answer requests. Deputy Manager and above.
Hand equipment to anyone Give items to any member of staff, not only their team. Senior Manager and Admin.
Add and edit equipment Add items, change details, build the categories. Admin.
See equipment reports The totals above. Manager and above.

So out of the box a deputy equips their own officers, a manager equips their whole team and reads the numbers, a senior manager reaches the entire company, and everyone else can ask for what they need and confirm what arrives. Customers on the portal never see equipment at all.

Switching parts of the app off

Admin → Features has a switch for each optional part of the helpdesk:

Feature What it covers
Equipment Laptops, phones and accessories: who holds what, requests, handovers and reports.
Documents Runbooks, guides and notes shared with roles, people or everyone.
Clients The outside companies you work with, and the issues raised for them.
Webhooks Signed messages to Slack, Teams or your own systems.
Analytics & reports Charts, team performance and the monthly report.

Switched off, a part of the app stops existing: it leaves the menu, its pages can't be opened even by typing the address, and the server answers "not found" for its API — so nothing half-works and nothing is left for someone to poke at. Nothing is deleted. Switch it back on and every item, document and setting is exactly where you left it.

Only roles with Turn parts of the app on and off can change this (the Admin role by default), and every change is written to the audit log with who made it.

Issues, people, roles, email and sign-in are the core of the helpdesk and cannot be switched off.

Single sign-on

Open Admin → Sign-in & SSO. Passwords always keep working, so nobody is locked out if the identity provider is down.

  1. Choose the type: OpenID Connect (most modern providers), SAML 2.0 (ADFS and other enterprise providers) or OAuth 2.0 (GitHub, or providers without OpenID Connect).
  2. Register the helpdesk with your provider. The page shows the exact address to copy (redirect URI, or for SAML the entity ID and ACS URL). Open the page at the address people really use, for example https://helpdesk.yourcompany.com, before you copy it.
  3. Paste the provider's details and press Check settings. The page contacts the provider (or reads the SAML metadata and certificate) and says what is wrong, if anything.
  4. New people: choose whether accounts are created on first sign-in (with the default role), and limit that to your email domains, for example yourcompany.com.
  5. Switch it on and Save. The login page now shows Sign in with …. Test it in a private window.

Recent sign-in problems at the bottom of the page lists failed attempts and why, which helps while you set it up.

Provider What to enter
Microsoft Entra ID OpenID Connect. Issuer https://login.microsoftonline.com/<tenant ID>/v2.0, the app's client ID and a client secret. Add the redirect URI as a Web platform.
Google Workspace OpenID Connect. Issuer https://accounts.google.com, a Web application OAuth client ID and secret.
Okta, OneLogin, JumpCloud OpenID Connect (issuer = your org URL) or SAML (paste the metadata XML).
Keycloak OpenID Connect with issuer https://<host>/realms/<realm>. For SAML, turn off Client signature required.
ADFS SAML: download the federation metadata XML and paste or upload it.

For SAML, the provider must sign the response or the assertion. Leave assertion encryption off, and send the email address as the Name ID or as an attribute called email.

Email

  • Outgoing mail (Admin → Outgoing email): choose Microsoft 365, Google or another SMTP server, then use Send test email. Microsoft 365 uses an app registration instead of a password: in Microsoft Entra, register an app, add the Office 365 Exchange Online application permission SMTP.SendAsApp, grant admin consent, create a client secret, and give the app access to the mailbox in Exchange Online. The page walks through it.
  • Email queues (Admin → Email queues): connect a mailbox (Microsoft 365 with IMAP.AccessAsApp, Google, or any IMAP server). New messages become issues. A reply to an issue email becomes a comment. Use Check connection before saving, and pause a queue with its switch.
  • What gets emailed, once outgoing mail is set up: the person who raised an issue hears when it is logged, when it is marked resolved (asking them to confirm), when it is closed, and when it is reopened; whoever is assigned an issue is told which one; and public replies go to the issue's contact. Each person chooses which of these they want under Settings -> Notifications. A mail problem never blocks the action that caused it.
  • The emails themselves share one clean design: your company name, a card with the reference, title, who raised it and the priority, a button straight to the issue, and a plain-text version for mail clients that want one. Admins can rewrite any of them under Admin -> Outgoing email -> Templates: the editor lists the placeholders, previews the result with example values, and Use the built-in design puts the helpdesk's own version back. Installs made before this design was added pick it up automatically.
  • Password reset emails and links in messages use the helpdesk's address from APP_URL (set by setup.sh).

Webhooks

Admin → Webhooks sends a message to another system when an issue is created, assigned, commented on or changes status. Slack, Microsoft Teams and Discord addresses get messages in their own format; anything else gets JSON. Use Send test to check it.

When you set a secret, every message carries X-Helpdesk-Timestamp and X-Helpdesk-Signature: sha256=<hex>, the HMAC-SHA256 of "<timestamp>.<body>" with the secret. Recompute it on your side and compare to be sure a message really came from the helpdesk.

Install on a fresh Ubuntu server

What you need

  • A server running Ubuntu 24.04 or 22.04 (Debian 12 also works). It can be completely empty: the installer installs Docker and everything else itself.
  • 2 GB of memory or more (4 GB recommended), 2 CPU cores, and 20 GB of disk.
  • SSH access as root, or as a user who can use sudo.
  • Optional: a domain name, if you want HTTPS (see below).

Step 1: copy the project to the server

On your Windows PC, open PowerShell in the folder that contains the project folder, then pack and upload it:

tar -czf helpdesk.tar.gz --exclude=.env --exclude=backups --exclude=node_modules --exclude=.next peppermint-main
scp helpdesk.tar.gz USER@SERVER_IP:~/

(The --exclude parts keep settings and files from your PC out of the upload, so they can never overwrite the server's own.)

On the server, unpack it into /opt/helpdesk:

ssh USER@SERVER_IP
sudo mkdir -p /opt/helpdesk
sudo tar -xzf ~/helpdesk.tar.gz -C /opt/helpdesk --strip-components=1

(With Git instead: sudo git clone YOUR_REPO_URL /opt/helpdesk. With WinSCP, just copy the folder over.)

Step 2: run the installer

cd /opt/helpdesk
sudo bash setup.sh

It asks four questions. Press Enter to accept a suggestion:

  1. Domain name: press Enter to use the server's IP address, or type a domain such as helpdesk.yourcompany.com to get HTTPS.
  2. Admin email for the first account.
  3. Admin password: press Enter to get a strong generated one.
  4. Company name shown in the app.

After that it runs on its own. The first install takes about 10–20 minutes, mostly building the app, and ends like this:

==================================================================
  Your helpdesk is live!
==================================================================

  Open:          http://203.0.113.10
  Admin email:   you@yourcompany.com
  Password:      ****************

Step 3: sign in and set it up

  1. Open the address and sign in with the admin account.
  2. Admin → Branding: upload your logo and check the company name.
  3. Admin → Issue settings: issue types, priorities with their deadlines, categories and required fields.
  4. Admin → Roles & permissions: check what each role can do and adjust it for your company.
  5. Admin → People: add people and give each a role. Then use the Team page to set who reports to whom and which staff member handles each user's issues.
  6. Admin → Outgoing email: connect a mailbox so people get email notifications and password resets.
  7. Optional: Admin → Sign-in & SSO so staff sign in with their company account (see Single sign-on), and Admin → Email queues to turn emails into issues.

What setup.sh does

  1. Checks the system, memory, free disk space and that all project files are present.
  2. Asks for the settings and saves them in .env, with a random database password and login secret. Only root can read the file.
  3. Installs a few basic tools and, on servers with less than about 3.5 GB of memory, adds a swap file so the build doesn't run out of memory.
  4. Installs Docker Engine and Docker Compose from Docker's official repository and starts Docker on boot.
  5. Opens ports 80 and 443 in the ufw firewall (if it's on) and checks nothing else is using them.
  6. Builds the app and starts three containers: db (PostgreSQL 18), app (the helpdesk) and proxy (Caddy, the web server that also handles HTTPS).
  7. Checks the site answers, then schedules a daily backup at 02:30.

Everything is logged to /var/log/helpdesk-setup.log. It is safe to run again: it keeps your settings and data, rebuilds the app and restarts it.

To install without any questions (for scripts), pass everything as options:

sudo bash setup.sh --yes --domain helpdesk.example.com --email you@example.com --company "Acme Ltd"

sudo bash setup.sh --help lists all options.

Install on a fresh Windows Server

Prefer Windows? There is an installer for it. Copy this folder onto the server, open the windows folder, then right-click Install Helpdesk.bat and choose Run as administrator.

It installs Node.js and PostgreSQL if they are missing, creates the database with a strong password, writes .env, builds the app, starts it, registers it to start again whenever Windows starts, opens the firewall, and prints the address and the first sign-in. Roughly 10-20 minutes on a clean server.

powershell -ExecutionPolicy Bypass -File windows\install.ps1 -DryRun    # check without changing anything
powershell -ExecutionPolicy Bypass -File windows\install.ps1 -Port 8080 # if 80 is taken
powershell -File windows\manage.ps1 status                              # is it running
powershell -File windows\manage.ps1 backup                              # save the database
powershell -File windows\manage.ps1 update                              # backup, rebuild, restart

Full details, including what to do when PostgreSQL is already installed and how to put it behind HTTPS, are in windows/README.md.

Use a domain name with HTTPS

  1. At your domain provider, create an A record, for example helpdesk.yourcompany.com → your server's public IP address.

  2. Make sure ports 80 and 443 are open to the internet, including in your cloud provider's firewall (it may be called a "security group").

  3. Run the installer with the domain. This works for a new install and for an existing one:

    cd /opt/helpdesk
    sudo bash setup.sh --domain helpdesk.yourcompany.com

Caddy gets a free certificate from Let's Encrypt, renews it automatically, and sends http:// visitors to https://.

  • Using Cloudflare? Set the record to DNS only (grey cloud) until the certificate has been issued. After that you can turn the orange cloud on with SSL mode Full (strict).
  • Going back to the IP address: sudo bash setup.sh --no-domain

Everyday commands

Run these inside /opt/helpdesk:

Command What it does
sudo ./manage.sh status Shows the three services, the web address, the latest backup and free disk space.
sudo ./manage.sh logs Live logs of everything (Ctrl+C to stop). logs app, logs db or logs proxy for just one.
sudo ./manage.sh stop Stops the helpdesk. Nothing is deleted.
sudo ./manage.sh start Starts it again. It also starts by itself after a reboot.
sudo ./manage.sh restart Restarts and applies any changes made in .env.
sudo ./manage.sh backup Makes a backup now.
sudo ./manage.sh restore FILE Replaces all data with a backup (see below).
sudo ./manage.sh update Backs up, rebuilds the app from the files in the folder, and restarts.
sudo ./manage.sh reset-password EMAIL Gives an account a new, generated password. Add a password after the email to choose it yourself. Without an email, it lists the admin accounts.

Backups and restore

  • Automatic: every night at 02:30 into /opt/helpdesk/backups/. The newest 14 are kept (change this with BACKUP_KEEP in .env).

  • Each backup is one file, helpdesk-YYYYMMDD-HHMMSS.tar.gz, holding the full database and all uploaded files.

  • Manual: sudo ./manage.sh backup

  • Copy backups off the server. A backup on the same disk won't help if the server itself is lost. From your PC:

    scp USER@SERVER_IP:/opt/helpdesk/backups/helpdesk-*.tar.gz .

    (The backups folder belongs to root. If your SSH user isn't root, first run sudo cp /opt/helpdesk/backups/*.tar.gz ~/ && sudo chown $USER ~/helpdesk-*.tar.gz on the server, then copy from ~/.)

Restore:

cd /opt/helpdesk
sudo ./manage.sh restore backups/helpdesk-20260915-023000.tar.gz

It asks you to type restore to confirm. It then saves the current data as a -before-restore backup, and loads the database in a single step: if anything goes wrong halfway, the database stays exactly as it was.

Updating

  1. Pack the new version on your PC with the same tar ... --exclude=.env ... command as in Step 1, upload it, and unpack it over the old one. The server's .env and backups/ are kept:

    sudo tar -xzf ~/helpdesk.tar.gz -C /opt/helpdesk --strip-components=1
  2. Rebuild and restart. This makes a backup first, and any database changes are applied automatically when the app starts:

    cd /opt/helpdesk
    sudo ./manage.sh update
  3. If a page looks old in your browser, press Ctrl+Shift+R once.

If you installed with Git, manage.sh update runs git pull for you first.

Moving to a new server

  1. On the old server, run sudo ./manage.sh backup and copy the new file from backups/ to your PC.
  2. On the new server, copy the project over and run sudo bash setup.sh (see Install).
  3. Copy the backup into /opt/helpdesk/backups/ on the new server and run sudo ./manage.sh restore backups/THE-FILE.tar.gz.

Everyone signs in with the same accounts and passwords as before.

Settings (.env)

setup.sh writes /opt/helpdesk/.env for you. .env.example explains each setting. After changing something, run sudo ./manage.sh restart.

Setting Meaning Default
SITE_ADDRESS A domain name (HTTPS), or :80 for plain HTTP on the server's IP address. :80
APP_URL The full address people open, used for links in emails and webhook messages. setup.sh fills it in from the web address. from setup
HTTP_PORT, HTTPS_PORT Ports on the server. Change them only if 80/443 are taken; HTTPS certificates need 80 and 443. 80, 443
ADMIN_EMAIL, ADMIN_PASSWORD, ADMIN_NAME The first admin account. Used only when the database is first created. from setup
COMPANY_NAME Company name at the first start. Change it later in Admin → Branding. Helpdesk
POSTGRES_PASSWORD Database password (random). Don't change it after installing: the database keeps the old one. random
SECRET Signs logins (random). Changing it signs everyone out. random
POSTGRES_USER, POSTGRES_DB Database user and database name. helpdesk
BACKUP_KEEP How many backups to keep. 14

Run it on your own PC

To try it out or develop on Windows with Docker Desktop, open PowerShell in the project folder and run:

docker compose -p peppermint -f docker-compose.local.yml up -d --build

Then open http://localhost:3000 and sign in with admin@admin.com / 1234. Change that password straight away. This file uses fixed test passwords, so never use it on a real server.

To try the real server setup on your PC instead, copy .env.example to .env, replace every change-me with a long random value (use only letters and digits for POSTGRES_PASSWORD), set HTTP_PORT=8080, then run docker compose up -d --build and open http://localhost:8080.

Troubleshooting

Problem What to do
$'\r': command not found or bad interpreter The files picked up Windows line endings. Run sed -i 's/\r$//' setup.sh manage.sh, then try again. (sudo bash setup.sh repairs this by itself.)
"Port 80 is already used by apache2" (or nginx) Stop that program with sudo systemctl disable --now apache2, or use other ports: sudo bash setup.sh --http-port 8080 --https-port 8443.
The build stops with "killed" or exit code 137 The server ran out of memory. Use a server with at least 2 GB (the installer adds swap automatically), then run setup.sh again.
A network error during the build Run sudo bash setup.sh again. Finished steps are cached, so it continues where it stopped.
The site doesn't open from your browser Allow ports 80/443 in your cloud provider's firewall, and check sudo ./manage.sh status.
HTTPS isn't working Check that the A record points to this server, that ports 80 and 443 are open, and that no AAAA (IPv6) record points somewhere else. Watch progress with sudo ./manage.sh logs proxy.
A page looks old after an update Press Ctrl+Shift+R.
Forgot the admin password sudo ./manage.sh reset-password you@yourcompany.com
Anything else sudo ./manage.sh logs app and /var/log/helpdesk-setup.log usually show the reason.

Security checklist

  • .env holds the secrets and only root can read it. Never share it or commit it to Git.
  • Only ports 80 and 443 are open to the outside. The database and the app's internal ports are not reachable from the internet.
  • For real use, set up a domain with HTTPS so passwords are never sent unencrypted.
  • Give everyone their own account with the lowest role they need, and remove accounts that are no longer used.
  • With single sign-on, limit automatic accounts to your own email domains (Admin → Sign-in & SSO → New people).
  • Built in: repeated wrong passwords lock the address for 15 minutes; signing out, a password change or a reset ends the other sessions straight away; reset codes expire after 30 minutes and lock after 5 wrong tries.
  • The sign-in is held in a cookie the browser will not give to page scripts (HttpOnly, SameSite=Lax, and Secure once APP_URL starts with https://), so a stray script can't walk off with someone's session. Document text is stored as editor blocks only, never as raw HTML.
  • Keep the server updated: sudo apt update && sudo apt upgrade. To install security updates automatically, run sudo apt install unattended-upgrades once.
  • Sign in to the server with SSH keys rather than passwords.
  • Keep copies of your backups somewhere other than the server.

How it is built

 browser ──► proxy (Caddy: ports 80/443, HTTPS certificates)
               │
               ▼
             app  (web app on port 3000 + API on port 5003, run by PM2)
               │
               ▼
             db   (PostgreSQL 18)

 Stored data (Docker volumes): pgdata = database, uploads = attached files,
                               caddy_data = HTTPS certificates
Path What it is
apps/client The web app (Next.js, React, Tailwind)
apps/api The API (Fastify, Prisma). Database migrations run automatically at start.
dockerfile Builds the app image
docker-compose.yml The server setup: db, app, proxy
docker-compose.local.yml Quick local testing on your PC
Caddyfile Web server and HTTPS settings
setup.sh The installer
manage.sh Everyday commands
.env.example Every setting, explained

License and credits

  • Based on Peppermint by Jack Andrews and contributors.
  • Licensed under the GNU AGPL v3; see the license file. In short: if people use a changed version of this software over a network, you must offer them its source code, including your changes.
  • Made with ❤️ by Abdul Manan: abdulmanan.tech · LinkedIn · GitHub