When creating telegraph page, by default requests.post is sending data with
'Content-type': 'application/x-www-form-urlencoded'
It's huge, because json payload is percent-encoded, for example this dict (43 bytes):
{"key": "value with non-ascii: абвгд"}
looks like this when sending request (97 bytes):
%7B%22key%22%3A%20%22value%20with%20non-ascii%3A%20%5Cu0430%5Cu0431%5Cu0432%5Cu0433%5Cu0434%22%7D
Telegra.ph accepts 'Content-Type': 'application/json' header and you can use this feature instead to avoid percent-encoding.
Use 'Content-Type': 'application/json' instead of default percent-encoding
Use json.dumps(data, ensure_ascii=False). Telegra.ph accepts utf-8 and no need to escape into \uXXXX for non-ascii characters.
Use separators without spaces json.dumps(data, ensure_ascii=False, separators=(',', ':'))
Also you don't need to separately convert content property to string, it can be sent like regular json
Unfortunately, this will not help to put more than 64 kb on a page, it helps only for reducing network overhead.
When creating telegraph page, by default requests.post is sending data with
'Content-type': 'application/x-www-form-urlencoded'It's huge, because json payload is percent-encoded, for example this dict (43 bytes):
{"key": "value with non-ascii: абвгд"}looks like this when sending request (97 bytes):%7B%22key%22%3A%20%22value%20with%20non-ascii%3A%20%5Cu0430%5Cu0431%5Cu0432%5Cu0433%5Cu0434%22%7DTelegra.ph accepts
'Content-Type': 'application/json'header and you can use this feature instead to avoid percent-encoding.'Content-Type': 'application/json'instead of default percent-encodingjson.dumps(data, ensure_ascii=False). Telegra.ph accepts utf-8 and no need to escape into\uXXXXfor non-ascii characters.json.dumps(data, ensure_ascii=False, separators=(',', ':'))Unfortunately, this will not help to put more than 64 kb on a page, it helps only for reducing network overhead.