It's a copy of Google Reader API.
The only thing you need to support BazQux Reader is to change endpoint URLs from
google.com to bazqux.com (or www.google.com to www.bazqux.com) in your code.
You need to have active account (either free trial or paid one) in BazQux Reader. If you're client app developer you could mail you account id to support@bazqux.com and I will mark your account as forever free.
If you haven't registered via email & password (logged in via Google/Facebook/Twitter/OpenID) you need to set username & password in settings (top-right corner) => Account.
API implementation is tested and works with Reeder, Feeddler, Web Subscriber and Vienna RSS. So there is a high probability it will work with your App without any hassle.
BazQux Reader also implements Fever API (local copy), which works in Unread and ReadKit.
BazQux Reader does not automatically mark items as read
after 30 days. So please don't add ot=CurrentTime-30days
when you get unread items
(s=user/-/state/com.google/reading-list&xt=user/-/state/com.google/read).
You may miss some unread items this way.
In general it's not right approach to use ot. Right way is to get ids of all, unread and starred items limited by n=<number_of_items> then remove items that are no longer exists, fetch new items and sync unread/starred state. Read more in the next section.
Please, DON'T "sync" by downloading each feed stream separately. User can have 3000 feeds and most of them won't be updated every hour. That's almost 3k of completely unnecessary requests.
Limiting amount of fetched items per feed with n option is very inefficient too. One feed can have 250 new items and another only 5 items from the whole last year. You either miss some articles or will download the same articles each "sync".
And, please, DON'T "sync" by downloading reading-list stream with ot option (or, even worse, download each feed with ot). It won't work!
Few reasons why using ot is a bad idea:
-
User subscribed to a new feed and it has posts before
otinreading-listthat you won't sync (please, do not sync 3000 feeds or sync this new feed separately to solve this). -
User marked some 5 years old item (or even yesterday one) as unread. Oh, it's not in the
reading-listor your feed filtered byotagain. -
BazQux Reader does not automatically mark items as read after 30 days. If you add
ot=CurrentTime-30daysyou will miss some items. -
BazQux Reader removes old items from feeds to keep 500 last items per feed (to not blow up the database). So if you sync only new items you may get into situation when you will have items (in high volume feeds) in your app that are no longer in BazQux Reader.
-
And BazQux Reader has filters. Real filters that really hide items (from both user and apps) instead of marking them as read. And user could change filters at any time. So any item from the past could be removed or added to the
reading-listat any time and your app will miss them. And you (and me) will get a support request. -
And what about starred items? User can unstar any item in the past. Please, don't download all starred items every time. Some users have >10k starred items.
How to do it right? Use item IDs, Luke!
General syncing process is:
-
Fetch list of feeds and folders.
-
Fetch list of tags (it contains folders too, so you need to remove folders found in previous call to get tags).
-
Fetch ids of unread items
(user can easily have 1000000 unread items so, please, add a limit on how many articles you sync, 25000 could be a good default, customizable limit is even better).
-
Fetch ids of starred items
(100k starred items are possible, so, please, limit them too, 10-25k limit is a good default)
-
Fetch tagged item ids by passing
s=user/-/label/TagNameparameter. -
Remove items that are no longer in unread/starred/tagged ids lists from your local database.
-
Fetch contents of items missing in database.
-
Mark/unmark items read/starred/tagged in you app comparing local state and ids you've got from the BazQux.
Use edit-tag to sync read/starred/tagged status from your app to BazQux.
- https://bazqux.com/accounts/ClientLogin
- https://bazqux.com/reader/api/0
- https://bazqux.com/reader/atom
Getting lists of all/unread items/ids in json/atom formats, marking items read/unread, starring, tagging, adding/removing/renaming of subscriptions, folders and tags (smart streams works like tags from API), custom subscriptions ordering -- everything is supported.
> curl https://bazqux.com/accounts/ClientLogin -d Email=foo -d Passwd=bar
Error=BadAuthentication
> curl https://bazqux.com/accounts/ClientLogin -d Email=realuser -d Passwd=***
SID=unused
LSID=unused
Auth=cltoken
Where cltoken is a client login token that you must pass to all other API calls in
Authorization header in form GoogleLogin auth=cltoken.
cltoken currently expires in 14 days. If you get 401 Unauthorized reply from any other request earlier, you need to relogin with existing credentials. Show user the login dialog with ability to enter new credentials only if relogin has failed.
You can test most API calls right in your browser when you signed in BazQux Reader.
This is an extension to Google Reader API. Since BazQux Reader is a commercial app login can fail even with correct login & password if free trial or subscription has expired. This information is passed in X-BQ-LoginErrorReason HTTP response header. There are two possible values YearSubscriptionExpired and FreeTrialExpired. And there are two test accounts for it:
curl -v https://bazqux.com/accounts/ClientLogin -d Email=se -d Passwd=1
...
< X-BQ-LoginErrorReason: YearSubscriptionExpired
...
curl -v https://bazqux.com/accounts/ClientLogin -d Email=fte -d Passwd=1
...
< X-BQ-LoginErrorReason: FreeTrialExpired
...
I strongly suggest you to check this header and display corresponding error messages "Login failed: Your subscription has expired" or "Login failed: your free trial has expired". This can greatly reduce a number of "why I can't login today?" support requests.
https://bazqux.com/reader/ping
> curl https://bazqux.com/reader/ping
Unauthorized
> curl https://bazqux.com/reader/ping -H "Authorization: GoogleLogin auth=cltoken"
OK
https://bazqux.com/reader/api/0/token
It's not used currently but may be used later the same way Google Reader uses it (expires in 30 minutes, with "x-reader-google-bad-token: true" header set).
> curl https://bazqux.com/reader/api/0/token -H "Authorization: GoogleLogin auth=cltoken"
Token123
https://bazqux.com/reader/directory/search
> curl https://bazqux.com/reader/directory/search?q=foo -H "Authorization: GoogleLogin auth=cltoken"
Search is not yet supported
https://bazqux.com/reader/api/0/user-info
> curl https://bazqux.com/reader/api/0/user-info -H "Authorization: GoogleLogin auth=cltoken"
{"userId":"01234567890123456789","userName":"realuser","userProfileId":"01234567890123456789","userEmail":"realuser",
"isBloggerUser":true,"signupTimeSec":1234567890,"isMultiLoginEnabled":true}
BazQux Reader uses dummy 01234567890123456789 user id for all users and accept any user in Google Reader labels
or states user/Abracadabra/label/MyFolder = user/01234567890123456789/label/MyFolder = user/-/label/MyFolder.
https://bazqux.com/reader/api/0/preference/list (?output=json)
The only preference is alphabetical sorting of subscriptions (custom ordering is not yet available on website).
https://bazqux.com/reader/api/0/friend/list (?output=json)
Empty list for compatibility.
https://bazqux.com/reader/api/0/preference/stream/list (?output=json)
Contains information about sorting (alphabetical) and expanded/collapsed state of folders.
sortId is just a number of a feed or folder for current user. The same number is also first half of
ItemId so beware that ItemIds are not unique between users
while unique for single user.
https://bazqux.com/reader/api/0/preference/stream/set
You may only set k=subscription-ordering&s=...&v=.... Other parameters are ignored.
https://bazqux.com/reader/api/0/tag/list (?output=json)
Only contains list of folders, tags, smart streams and some google specific feeds.
https://bazqux.com/reader/api/0/subscription/list (?output=json)
Always contain htmlUrl not depending on favicons setting. firstitemmsec is always dummy 1234567890000 (it seems
that no one use it).
https://bazqux.com/reader/subscriptions/export
https://bazqux.com/reader/api/0/subscription/quickadd?quickadd=xkcd.com (?output=xml)
BazQux Reader currently doesn't differ GET/POST and query string or form parameters. But you must use POST and append T=token to be future proof (and to be compatible with Google Reader API).
https://bazqux.com/reader/api/0/subscription/edit
...?ac=subscribe&s=feed/xkcd.com&t=XKCD
...?ac=unsubscribe&s=feed/http://xkcd.com/atom.xml
You can put one subscription in many folders.
https://bazqux.com/reader/api/0/unread-count (?output=json)
https://bazqux.com/reader/api/0/stream/items/ids (?output=json)
s=user/-/state/com.google/reading-list
s=user/-/state/com.google/starred
s=user/-/state/com.google/read
s=user/-/state/com.google/broadcast - empty results
s=user/-/state/com.google/created - empty results
s=user/-/label/... - folder, tag or smart stream
s=feed/...
r=o - oldest first ranking.
xt=... - everything possible in s=.
it=... - only messages with specific tags (starred items in feed). NB: It's an extension to GR API.
ck=... - ignored.
ot=... - minimum download time in seconds. It's time after which item appeared in reader, not the published time. Please don't add it when you get unread items list (there are no 30 days limit on a number of unread items). Internally 180 seconds are subtracted from ot to take caching into account (since previous call may not return last few items due to caching). So you can get few items before ot. Please, do not use ot.
nt=... - maximum download time in seconds.
ts=... - maximum published time in μs (1e-6 seconds). This could be used in mark-all-as-read call to implement "Mark older than N days" feature.
c=... - continuation from previous request (it's just an item id and hence never expire).
n=... - number of items. Default is 20, maximum is 50000.
includeAllDirectStreamIds=true - add source feed, tags and smart streams for each item.
Note that item ids are unique for one user but can overlap between users. Please use separate database for each account.
https://bazqux.com/reader/api/0/stream/items/contents (?output=atom)
Pass all your item ids in i= parameters (like i=item_id1&i=item_id_2&...&i=item_id_N) in HTTP POST request. Although it's possible to add them to URL I recommend POST to not bump into URL length limit.
No more than 1000 items at once.
I suggest to fetch 50-100 items per call on mobile and 100-250 items per call on desktop. Fetching 1000 items at once may be a bit faster but requires more memory on mobile and syncing progress is less visible.
Item ids are 64 bit signed integers. To be compatible with Google Reader they are sometimes represented in long form:
tag:google.com,2005:reader/item/<unsigned zero-padded 64 bit hexadecimal number>
For example
tag:google.com,2005:reader/item/000088960000047a = 150177826473082
tag:google.com,2005:reader/item/80484b00000e8003 = -9203023375158575101
All API calls accept both item ids formats. Note that /stream/items/ids return ids in short form but /stream/items/contents in long form. You can safely convert them to 64 bit signed integer.
More here https://raw.githubusercontent.com/mihaip/google-reader-api/master/wiki/ItemId.wiki
https://bazqux.com/reader/api/0/stream/contents (?output=atom)
https://bazqux.com/reader/atom (= /stream/contents?output=atom)
The same options as in stream/items/ids are supported. When no subscription given defaults to reading-list.
n=1000 maximum.
https://bazqux.com/reader/api/0/edit-tag
i= - list of item ids (as in /stream/items/contents).
a= - add tag.
r= - remove tag.
Tags are: user/-/state/com.google/read for marking read/unread, user/-/state/com.google/starred for starring
and user/-/label/... for tagging.
No more than 10000 items to tag at once.
https://bazqux.com/reader/api/0/mark-all-as-read
s= - feed you want to mark as read (you could use all parameters that /stream/items/ids accepts).
No more than 50000 items are marked at once.
https://bazqux.com/reader/api/0/rename-tag?s=user/-/label/Comics&dest=user/-/label/NiceComics
https://bazqux.com/reader/api/0/disable-tag?s=user/-/label/NiceComics
Feeds are not removed, only folder tag is (like in Google Reader).
NB: This method is an extension to GR API.
https://bazqux.com/reader/api/0/import/opml
Post OPML data in opml parameter (in usual application/x-www-form-urlencoded form). It will return percent of feeds currently processed (integer number in plain text).
You can then poll (each 3 seconds for example)
https://bazqux.com/reader/api/0/import/percent-complete
till it return "100".
Description of API from Mihai Parparita (Google Reader project manager), I suggest you to look here first when you need details about API calls:
https://code.google.com/p/google-reader-api/w/list
Docs from Nick Bradbury (NetNewsWire developer):
http://inessential.com/2013/03/14/google_reader_api_documentation
One more descrption: