Skip to content


Switch branches/tags

Name already in use

A tag already exists with the provided branch name. Many Git commands accept both tag and branch names, so creating this branch may cause unexpected behavior. Are you sure you want to create this branch?


Failed to load latest commit information.
Latest commit message
Commit time
May 28, 2013 00:18
September 25, 2020 15:17
July 30, 2021 21:35

BazQux Reader API

It's a copy of Google Reader API.

The only thing you need to support BazQux Reader is to change endpoint URLs from to (or to 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 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/ 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.

The Right Way to Sync

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 ot in reading-list that 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-list or your feed filtered by ot again.

  • BazQux Reader does not automatically mark items as read after 30 days. If you add ot=CurrentTime-30days you 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-list at 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:

Use edit-tag to sync read/starred/tagged status from your app to BazQux.

Main API endpoints

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 -d Email=foo -d Passwd=bar

> curl -d Email=realuser -d Passwd=***

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.

Login error details

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 -d Email=se -d Passwd=1
< X-BQ-LoginErrorReason: YearSubscriptionExpired

curl -v -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.


> curl

> curl -H "Authorization: GoogleLogin auth=cltoken"


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 -H "Authorization: GoogleLogin auth=cltoken"

Directory search

> curl -H "Authorization: GoogleLogin auth=cltoken"
Search is not yet supported

User info

> curl -H "Authorization: GoogleLogin auth=cltoken"

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.

Preferences list (?output=json)

The only preference is alphabetical sorting of subscriptions (custom ordering is not yet available on website).

Friend list (?output=json)

Empty list for compatibility.

Stream preferences 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.

Set stream preferences

You may only set k=subscription-ordering&s=...&v=.... Other parameters are ignored.

Tag list (?output=json)

Only contains list of folders, tags, smart streams and some google specific feeds.

Subscriptions list (?output=json)

Always contain htmlUrl not depending on favicons setting. firstitemmsec is always dummy 1234567890000 (it seems that no one use it).

Subscriptions OPML

Adding subscription (?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).

Editing subscription




You can put one subscription in many folders.

Unread count (?output=json)

Item ids (?output=json)




s=user/-/state/ - empty results

s=user/-/state/ - empty results

s=user/-/label/... - folder, tag or smart stream


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.

Fetching individual items (?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.

About item ids

Item ids are 64 bit signed integers. To be compatible with Google Reader they are sometimes represented in long form:,2005:reader/item/<unsigned zero-padded 64 bit hexadecimal number>

For example,2005:reader/item/000088960000047a = 150177826473082,2005: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

Fetching streams (?output=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.

Tagging items

i= - list of item ids (as in /stream/items/contents).

a= - add tag.

r= - remove tag.

Tags are: user/-/state/ for marking read/unread, user/-/state/ for starring and user/-/label/... for tagging.

No more than 10000 items to tag at once.

Marking 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.

Folder/tag/smart stream renaming

Folder/tag/smart stream removing

Feeds are not removed, only folder tag is (like in Google Reader).

Subscriptions import

NB: This method is an extension to GR API.

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)

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:

Docs from Nick Bradbury (NetNewsWire developer):

One more descrption:


BazQux Reader API Documentation






No releases published


No packages published