Migration
Migrate from Zendesk without losing your history
Your tickets take their place here with their original numbers and dates. We run the migration ourselves, and we tell you in advance what cannot come along — because a migration is judged on what you are not told, not on what you are promised.
The process
Five steps, two weeks as a rule
How long it takes depends mostly on the size of your history and on how long Zendesk takes to deliver the export. You will not write a line of code.
- 1
You give us read access
A read-only Zendesk API token is enough, or the archive from your export if you would rather not open anything up. We need no access to your production system.
- 2
We take stock of what is there
Ticket volume, custom fields, brands, groups, triggers, articles. You receive a written inventory and the list of what will not be able to come along, before you commit.
- 3
Dry-run migration
We migrate a representative sample into a trial workspace. You open it, you search for your tickets, you check your threads. Nothing is settled until you agree.
- 4
Full migration
The entire history is migrated while you carry on working in Zendesk. Tickets keep their numbers and their created, first-reply and resolved dates.
- 5
Address switchover
On the day, your support addresses point to Open HelpDesk. Open tickets are migrated last so that no message is lost during the transition.
Fidelity
What carries over, what changes, what is lost
No migration between two different products is a hundred per cent faithful. Here is ours, in detail — you will read it before you sign, not after.
Migrated as-is
- Tickets, with their Zendesk number and their real dates
- Conversation threads, every message attributed to its author
- Internal notes, kept distinct from public replies
- Contacts, organisations and how they are linked
- Attachments up to 10 MB
- Tags and compatible custom fields
- Knowledge base articles and their categories
Migrated, but changed
- Your custom statuses fall back to the six standard statuses
- Satisfaction rated 1 to 5 becomes satisfied or dissatisfied
- Tickets from phone, chat or social media are marked “email”
- Regular expression and decimal fields become text
- Article HTML is cleaned up for our editor
What does not come along
- Apps installed from the Zendesk marketplace
- Triggers and automations, to be rebuilt here — we do it with you
- Side conversations
- Attachments over 10 MB, listed separately so that you decide
- Requesters with no email address, which our model cannot represent
Over at Zendesk
Getting your data out of Zendesk
You can do it yourself, and you should: this data belongs to you, whether you come to us or not. Four things the official documentation puts badly.
The API works on every plan, the native export does not
The built-in export tools are not available on the Team plan. The REST API, on the other hand, answers whatever your plan is: that is the reliable route.
Take the JSON, not the CSV
The CSV only gives you an outline of the tickets. The JSON carries the tickets, the users, the organisations and above all the comments — which is to say your conversations.
Reckon with the rate limit
The incremental export API is capped at ten requests per minute. On a sizeable history, extraction is counted in hours: you start it and let it run.
Attachments come out as links, not as files
The export gives you URLs, not documents. You have to download them one by one, authenticated, before your access is cut off. It is the step everyone forgets and then regrets.
The big day
Switching your addresses without losing a message
This is the only delicate moment in a migration, and it can be prepared for. The method we apply:
You set up a forward from your support address to the address Open HelpDesk provides, keeping Zendesk running.
For a few days, both products receive mail. Your agents reply from Open HelpDesk and check that nothing is missing.
The tickets still open in Zendesk are migrated last, once they have gone quiet, so that no reply falls between the two.
When nothing more arrives in Zendesk, you change your MX records and close the account.
Reciprocity
And if you want to leave
The question deserves an honest answer rather than an argument. On the hosted version, a button in your settings downloads your whole history: the tickets with their numbers and their dates, the conversations in full, the internal notes, the contacts and the organisations, in a format any product can read. You have nothing to ask us for. And since the core is under AGPL-3.0, you can also install it on your own servers and take the whole database with a plain pg_dump.
And if I want to do it myself?
You can. An import screen lives in your workspace settings. You drop your Zendesk export there, you first run a dry-run migration that writes nothing and hands you the full report, then you launch the migration itself. We remain on hand for bulky or awkward histories — but you do not need us to get started.
Questions
What we get asked most often
- How much does the migration cost?
- Nothing. The migration is included whatever your plan, including if you stay on the free three-agent plan.
- How long does it take?
- Two weeks as a rule, half of it spent waiting on Zendesk. A history of several hundred thousand tickets takes longer: we say so from the inventory onwards.
- Are my ticket numbers kept?
- Yes. A ticket 48210 in Zendesk stays 48210 here, and your customers find their references again. New tickets pick the numbering back up above the highest number migrated.
- Do my customers receive emails during the migration?
- No, none at all. The migration triggers no rules, no acknowledgements and no SLA counters on the history — which is precisely why we write to the database rather than through the API.
- Can we migrate only part of it?
- Yes. Plenty of teams migrate only the last two years, or only the unresolved tickets, and keep their Zendesk export as an archive. We filter by date, by group or by brand.
- What becomes of our triggers and macros?
- Macros carry over. Zendesk triggers do not share the grammar of our rules: we rebuild them with you during the dry-run migration, and you check them on the sample.
- Does our data stay in Europe?
- Yes, on the hosted version: the servers are in France, and that is the default, not a billed option. Self-hosted, it never leaves your machines.
Let's talk about your history
Create a free workspace to try the product, or write to us with the approximate volume of your tickets: we reply with an inventory and a timeline.