Frappe Helpdesk implementation: Frappe Cloud vs self-hosting, setup and email, SLAs and go-live
Short answer
Frappe Helpdesk is an open-source help desk (AGPL-3.0) with no per-agent pricing — Frappe's own pages say unlimited agents, tickets and customers. You pay for hosting, not seats. On Nov 6, 2026 the Frappe Helpdesk pricing page showed Frappe Cloud site plans from $5 / ₹410 per month and server plans from $20 / ₹1,800 (with a higher server tier at $60 / ₹5,400). The Frappe Cloud pricing page describes the same site-vs-server split, with tooltips noting Hetzner site plans from $5 / ₹410 and AWS/OCI site plans from $25 / ₹2,050, and product-warranty messaging that differs slightly from the Helpdesk pricing FAQ (see below). Self-hosting is free via Frappe's easy-install script or Bench, but you own upgrades, backups and uptime. A sound implementation is: pick Cloud vs self-host deliberately, connect email correctly, design SLA policies with working hours and holidays, then go live in phases — portal and intake first, automations and deep customisation second.
Contents
- Who this guide is for
- Hosting prices and the no-per-agent model (checked Nov 6, 2026)
- Frappe Cloud vs self-hosting: how to choose
- Foundation setup before you invite agents
- Email setup that will not embarrass you on day one
- SLAs, working hours and holidays
- Phased go-live (no fake day counts)
- Operational pitfalls we see after go-live
- Best for / wrong choice when
- Frequently asked questions
Who this guide is for
This guide is for support and IT leaders who have chosen — or are close to choosing — Frappe Helpdesk and need an implementation path that does not pretend a greenfield help desk is a one-afternoon job. It is also for teams comparing Frappe Cloud hosting costs with self-hosting, and for operators already running ERPNext who want Helpdesk on the same bench.
Product capabilities and install steps come from Frappe's Frappe Helpdesk product page, Frappe Helpdesk installation docs, Frappe Helpdesk email communication docs and Frappe Helpdesk SLA docs, plus the frappe/helpdesk on GitHub repository, all checked on Nov 6, 2026. Hosting prices come only from Frappe's official pricing pages, checked the same day. We do not invent day-count project plans; phases below are sequential, not calendared.
Two official pricing pages, slightly different emphasis
Frappe publishes Helpdesk-oriented pricing at frappe.io/helpdesk/pricing and platform Cloud pricing at frappe.io/cloud/pricing. Both were checked on Nov 6, 2026. Headline entry prices align on a $5 starting point for small site hosting, but the Cloud page's plan tooltips distinguish Hetzner from $5 / ₹410 and AWS/OCI from $25 / ₹2,050, while the Helpdesk pricing FAQ says product support under Product Warranty applies on the $50/month plan or higher. The Helpdesk page also shows server starting prices of $20 and $60. When a figure matters to your order form, trust the live Frappe Cloud checkout for your region and plan, and treat marketing summaries as directional.
Hosting prices and the no-per-agent model (checked Nov 6, 2026)
Frappe's product and pricing pages are explicit: Helpdesk does not charge per agent. The Frappe Helpdesk pricing page FAQ states there are no limits on tickets, agents or customers based on pricing. You subscribe to Frappe Cloud compute (or you self-host).
| Option | What Frappe published (Nov 6, 2026) | Good starting point when | You should rethink when |
|---|---|---|---|
| Frappe Cloud — Sites | Helpdesk pricing page: from $5 / ₹410 for a single site on shared servers; Cloud page tooltips: Hetzner from $5 / ₹410, AWS/OCI from $25 / ₹2,050 | Trying Helpdesk, small team, one site, multiple Frappe apps | You need dedicated isolation, many sites, or stricter reliability guarantees |
| Frappe Cloud — Servers | Helpdesk pricing page: from $20 / ₹1,800 shared/dedicated, and $60 / ₹5,400 for superior dedicated isolation / advanced features | Production workloads, multiple sites/benches, predictable performance | You do not have budget for a VM-class plan and can accept shared site limits |
| Self-host | $0 licence (AGPL-3.0); you pay your own infrastructure and labour | You have Linux/Docker skills, compliance needs, or existing Bench/ERPNext | Nobody owns upgrades, monitoring, backups and restore tests |
- 14-day trial. Frappe's Helpdesk pricing FAQ says the free trial gives 14 days of access; after that you subscribe to a Frappe Cloud plan to keep the site.
- Product Warranty vs hosting support. All Helpdesk plans on Frappe Cloud include hosting-related support, per the FAQ. Product Warranty (bug fixes from Frappe engineers on first-party apps) is described as applying on the $50/month plan or higher on the Helpdesk pricing FAQ. Confirm on your chosen Cloud plan what warranty and Priority SLA options are attached.
- Education / non-profit. The same FAQ says discounts exist; you contact Frappe to apply.
- Unlimited agents is not unlimited operations. Seat cost is not your constraint — email deliverability, SLA design, agent training and knowledge quality still are.
For a feature-level comparison against Zendesk and Freshdesk, see our Frappe Helpdesk vs Zendesk vs Freshdesk guide. This article stays on implementation.
Frappe Cloud vs self-hosting: how to choose
Frappe's installation docs recommend trying Frappe Cloud before self-hosting. The Cloud pricing page also publishes a comparison table that contrasts UI-based deployments, automated backups, monitoring, disaster recovery and hosting support on Cloud against manual equivalents when you self-host. That table is marketing, but the underlying trade-off is real.
| Topic | Frappe Cloud | Self-host (easy-install / Bench / Docker) |
|---|---|---|
| Time to first site | Minutes; Marketplace install if you already have a Cloud site | Official easy-install production path is measured in minutes on a clean host, but DNS, TLS and mail still take care |
| Upgrades | Scheduled/managed from the Cloud UI (per Frappe's comparison) | You pull app updates, run migrations, and fix breaks |
| Backups / monitoring | Positioned as automated on Cloud | You design backup, offsite copy, restore tests, alerts |
| Compliance / data residency | Choose provider/region options Frappe Cloud offers (multi-cloud messaging on the pricing page) | Full control — and full responsibility |
| Cost shape | Monthly Cloud subscription in USD/INR as published | Infrastructure invoice + engineer time; often higher total if you value labour honestly |
| Exit | AGPL code means you can leave; plan export and DNS cutover | You already have the keys; still plan app and DB backup discipline |
Our view
Most support teams without an existing Frappe/ERPNext platform team should start on Frappe Cloud. Self-host when you already run Bench in production, when a regulator or customer contract forces it, or when Cloud plan economics at your scale clearly lose to a machine you already operate. Do not self-host to "save $5" if nobody is on-call for restore drills.
Official install paths (self-host)
Per the Frappe Helpdesk installation docs:
- Managed: one-click on Frappe Cloud, or install from the Frappe Cloud Marketplace onto an existing site.
- Production self-host: download
easy-install.pyfrom frappe.io and runpython3 ./easy-install.py deploywith project, email, imageghcr.io/frappe/helpdesk, versionstable, apphelpdesk, and your sitename. Frappe says the setup takes about five minutes for the platform piece. - Existing Bench / ERPNext site:
bench get-app helpdesk --branch mainthenbench --site your_site_name install-app helpdesk. - Development: follow the GitHub repository instructions rather than the production easy-install path.
Foundation setup before you invite agents
Whether Cloud or self-host, lock these foundations before ticket volume arrives:
Foundation checklist
- Create the company/site name, default timezone, and working calendar you will use for SLAs.
- Define agent groups that match how work is actually triaged (for example Billing, Technical, Onboarding) — not your org chart fantasy.
- Add custom fields you know you need on day one (order ID, customer tier) and resist boiling the ocean.
- Create canned responses for the ten questions you already answer daily.
- Publish a small public knowledge base before go-live so the portal is not an empty room.
- Decide roles/permissions: who can delete tickets, export data, manage SLAs, and change email accounts.
Frappe positions Helpdesk as highly customisable on the Frappe Framework. That is a strength and a risk. Custom fields and assignment rules are cheap to add later; undoing a messy permission model after agents invent workarounds is not. Prefer a thin day-one configuration that matches current email triage, then iterate.
Email setup that will not embarrass you on day one
Email is still how most customers write in. Frappe's Frappe Helpdesk email communication docs describes replying from the ticket page, canned responses, and Email Account setup. Key points from that documentation:
- You can manage multiple incoming and outgoing Email Accounts; there must be at least one default outgoing and one default incoming account.
- On Frappe Helpdesk Cloud, the default outgoing email is set by Frappe.
- Email Accounts are created from the Email Account master; an Email Domain is required unless you use a provider Frappe lists as skippable.
- Save verifies credentials; Gmail-class providers may need 2FA and app passwords.
- Options exist for alternate login IDs, awaiting password, and ASCII password encoding.
Implementation practice (ours)
- Use a real support address you already own (support@yourdomain), with DNS you control. Frappe's product page states you can connect an existing support address so customers do not change how they reach you.
- Separate ingest from personal mailboxes. Agents should work inside Helpdesk, not from a shared Gmail UI in parallel, or you will double-reply.
- Test the full loop: external send → ticket create → agent reply → customer receives from the right From address → reply-all returns to the same ticket.
- Plan SPF/DKIM/DMARC for your sending domain before marketing go-live. Deliverability failures look like "the tool is broken".
- Document who may add Email Accounts. A second uncontrolled inbox connected to production is a common self-inflicted outage.
Cloud vs self-host mail
On Frappe Cloud, follow Frappe's default outgoing behaviour and their current guidance for custom domains. On self-host, you own the MTA or transactional provider. Do not go live on self-host SMTP that only works on your laptop VPN.
SLAs, working hours and holidays
Frappe's Frappe Helpdesk SLA docs is unusually clear about how timers work. Read it before you promise customers a one-hour response time you cannot meet across weekends.
Each SLA is linked to the HD Ticket doctype and includes default priority, enabled flag, optional valid-from/to dates, per-priority first response and resolution times, statuses that fulfil the SLA (for example Resolved, Closed), statuses that pause the timer (for example Replied), working hours, and an optional holiday list.
| Concept | What Frappe's docs say | Why implementers care |
|---|---|---|
| Priority table | First response and resolution times per priority | A policy that omits a priority will not apply to tickets on that priority |
| Paused on Replied | Resolution timer pauses while waiting on the customer | Stops punishing agents for customer silence |
| Working hours | Timers respect configured workdays/hours | Overnight tickets do not burn SLA while the team sleeps |
| Holiday list | Linked holidays pause timers; recurring weekends supported | Festival weeks stop generating false breaches |
| Assignment conditions | Conditions on ticket fields + priority membership | Billing bugs can get a tighter SLA than general how-tos |
| Rank | Lower rank wins among matches; unranked (0) applies last | Prevents a new catch-all from stealing tickets from a careful policy |
| Default SLA | Optional; considered last; still needs matching priority | Safety net — or omit it deliberately if unmatched tickets should have no SLA |
Design rules we use
- Write SLAs from channels you can staff, not from a competitor's marketing site.
- Start with one default policy and one tighter policy for a single VIP or revenue-critical condition — not eight policies on day one.
- Align holiday lists with HR's published calendar; wrong holidays create either false breaches or silent customer pain.
- Train agents on the Replied pause: a vague "checking" reply that leaves the customer without a question can pause timers incorrectly from a customer-experience perspective even when the software behaves as designed.
Phased go-live (no fake day counts)
We deliberately do not publish "go live in 14 days" claims. Volume, email complexity, and whether you migrate history all change the calendar. Use phases with exit criteria instead.
Phase A — Foundations
Site live on Cloud or self-host. Agents and groups created. Email loop tested with external addresses. Knowledge base has answers to the top repetitive questions. SLA policy with working hours exists. Success: internal pilot tickets only.
Phase B — Soft launch
A subset of real customers or a single region routes to Helpdesk. Assignment rules on. Canned responses in use. Daily stand-up on breaches and misroutes. Success: stable intake for the pilot segment; no parallel inbox for that segment.
Phase C — Cutover
DNS/mail routing for the remaining addresses. Old tool set to read-only or auto-reply with the new portal link. Open-ticket war room for the first business period after cutover. Success: old tool no longer accepts new work.
Phase D — Harden
Automations, split/merge norms, telephony if in scope, ERPNext customer sync if you use it, permission audit, backup/restore drill (especially self-host). Success: documented operating model owned by a named admin.
Migration notes
Frappe's product FAQ says you can bring existing support data into Helpdesk. Treat history migration as optional for go-live unless compliance requires it. Many teams go live with clean open tickets only, and archive the legacy tool read-only. If you do migrate, map contacts and ticket subjects carefully, and accept that workflows, macros and permissions will be rebuilt — not exported as living automations — the same pattern we describe in other help-desk migrations.
Go-live readiness
- Email round-trip proven from outside your network.
- SLA timers sampled against working hours and a holiday.
- Agents trained on Replied vs Resolved.
- Portal link and knowledge base linked from your website or intranet.
- On-call owner named for the first week of production mail.
- Export/backup path verified (Cloud UI or your self-host backup).
Operational pitfalls we see after go-live
Most Frappe Helpdesk projects that struggle do not fail on install. They fail on operating habits after the first real week of mail.
| Pitfall | What happens | Prevention |
|---|---|---|
| Parallel inbox remains open | Agents reply from Gmail and Helpdesk; customers get duplicates; SLAs lie | Hard cut of the soft-launch segment; monitor the old mailbox for a short drain period only |
| SLA without holidays | Festival weeks create mass breaches or agents ignore the timer | Link the holiday list before soft launch; sample a ticket across a weekend |
| Over-customisation in week one | Custom fields and assignment rules nobody uses; permissions nobody understands | Freeze non-critical customisation until Phase D |
| Self-host without restore drill | Backup exists on paper; restore fails when you need it | Restore to a scratch site once before calling production ready |
| Unlimited agents, unclear ownership | Everyone is an agent; nobody owns triage | Name a queue owner per group; measure first response by group |
Frappe's product page emphasises assignment rules, SLAs, knowledge base, dashboards, automations, split and merge, permissions, telephony and ERPNext customer sync. Those capabilities pay off only when someone owns them. Put a named admin on the weekly review of overdue tickets, failed email sync and SLA breaches — fifteen minutes, every week, from soft launch onward.
If you are comparing Frappe with Zendesk or Freshdesk on features rather than implementation, stop here and read the comparison guide first. If you have already chosen Frappe, the tables above are your operating agenda.
Best for / wrong choice when
Frappe Helpdesk implementation is a good fit when
- Per-agent SaaS pricing is distorting your team shape, and you want unlimited agents on a hosting bill.
- You are comfortable with an open-source product and Frappe's ecosystem (or already run ERPNext).
- You can own configuration (fields, SLAs, permissions) without needing a heavily packaged ITIL suite on day one.
- You want the option to self-host later without rewriting the application.
Wrong choice when
- Wrong: picking self-host with no platform owner. The licence is free; reliability is not.
- Wrong: expecting Zendesk-class marketplace depth on day one. Compare ecosystems honestly in our comparison guide.
- Wrong: skipping email authentication and blaming the app for spam folders.
- Wrong: copying another vendor's SLA matrix without staffing it. Frappe will measure the promises you configure.
Frequently asked questions
Does Frappe Helpdesk charge per agent?
No. Frappe's Helpdesk pricing FAQ states there are no limits on tickets, agents or customers based on pricing. You pay for Frappe Cloud hosting (or you self-host).
How much does Frappe Cloud cost for Helpdesk?
Checked on Nov 6, 2026: the Helpdesk pricing page showed site hosting from $5 / ₹410 and server plans from $20 / ₹1,800 and $60 / ₹5,400. The Cloud pricing page tooltips also note Hetzner site plans from $5 / ₹410 and AWS/OCI from $25 / ₹2,050. Confirm the live checkout for your region and plan.
Is there a discrepancy between helpdesk/pricing and cloud/pricing?
Both pages agree on a low $5 entry point for small site hosting, but they emphasise different plan ladders (Helpdesk page server $20/$60 cards; Cloud page provider-specific site minimums and warranty options). The Helpdesk FAQ's "$50/month or higher" Product Warranty note may not match every Cloud card you see — verify warranty on the exact plan you buy.
Can I self-host for free?
Yes. Frappe Helpdesk is AGPL-3.0 on GitHub. Official docs provide easy-install production deploy, Bench install on an existing site, and a development path. Infrastructure and operations cost remain yours.
How do SLAs handle nights and holidays?
Per Frappe's SLA documentation, timers respect working hours and linked holiday lists, and can pause on statuses such as Replied. First response and resolution due times are calculated with those factors.
How do I connect email?
Use Email Account records as described in Frappe's email documentation: default incoming and outgoing accounts, domain or supported provider, verified credentials. Test a full external round-trip before go-live.