How to get all jettons? And what's changing in TONAPI?
In a nutshell
There are two main ways: TONAPI (the /v2/jettons method) and Tonconsole.com (CSV export).
When using the /v2/jettons method for pagination, it's recommended to use the cursor — the last_account_id parameter, which is taken from the previous response. It looks something like this:

% curl -s 'https://tonapi.io/v2/jettons?limit=10' | jq -r '.jettons | last | .metadata.address'
0:249936139b63719c0afc44f883f4d8c15fd53bc6f93b6244debcf92b3e6c3ce1
% curl -s 'https://tonapi.io/v2/jettons?limit=10&last_account_id=0:249936139b63719c0afc44f883f4d8c15fd53bc6f93b6244debcf92b3e6c3ce1' | jq -r '.jettons | last | .metadata.address'
0:31bb2214dca832dd8091d90adc313e9834782934263764810cb28fc7f8fb3878

In more detail

1. The /v2/jettons method allows you to get real-time data.

At the moment there are almost 1 million jettons on TON. One request lets you get 1000 data rows. Theoretically, on the Lite tier (10 RPS), you could download all jettons in about a minute and a half.

But this method had a problem — the offset parameter. Until recently, this method's execution time grew proportionally to the offset parameter. To maintain compatibility and ensure high performance, we carried out an optimization that uses a cached jetton list when offset is used. This list is updated once a day.

At the same time, we've added a new parameter — last_account_id. It provides a more correct and efficient way to paginate to the most up-to-date data.

2. For large exports like this, it may be more convenient to use Tonconsole.

For example, a query like SELECT d.account_id FROM getmethods.get_jetton_data d lets you get the full list of jettons. In Tonconsole you can enable scheduled export runs. The "Download CSV" link is stable, and you can quickly get a ready-made dataset with a single request from the backend (for the query above, that's 60 MB).

Feel free to ask question in comments ✍️