Hacker News new | past | comments | ask | show | jobs | submit login
Introducing the premium full-archive search endpoint (blog.twitter.com)
106 points by jonbaer on Feb 1, 2018 | hide | past | favorite | 21 comments



Premium searches cost about $1 per query: https://developer.twitter.com/en/pricing/search-fullarchive


It's strange that page mentions rate limiting of 60 requests/min, 10 requests/sec. Are you sure the pricing is 100 requests per month?


If you go to their "packages" page[1] and select "premium", it gives you a table with "requests (per period)" and "month-to-month (price per period), and it is indeed "up to 100" for $99.00. For up to 500 tweets per request.

Seems pretty clear, though I can't fathom why they chose "periods" rather than "months". Crazy expensive.

[1]: https://developer.twitter.com/en/pricing/search-fullarchive

---

edit: though the more I read it, the less confident I become. wtf is up with this copy.


It sure seems that way. If I understand their copy correctly this looks really really really expensive. I'm assuming the tweets per request is the limit to the result set returned, which is crazy small.


They don't specifically define the "period". No matter what the period is the price per search goes from $1 to $0.76 depending on volume.

I wonder how they can justify these prices?


I could certainly see a "search of everything by user" be very much worth a $1 for groups like intelligence agencies or background check companies.

That cost is then passed back on to the requester.


> I wonder how they can justify these prices

I mean, you've got to think someone has done an expected demand / price model and decided that this will maximize profits. Only people they need to justify to are shareholders.


If there is one thing I despise about tech companies nowadays it is how everything always reduces down to advertising or marketing analytics. They are turning marketing people into some of the most obnoxiously demanding folks out there


Nobody is willing to pay for any of that stuff directly so advertising is the only way to make money these days.


Then at this point I would rather all the nonfree stuff just go away. (Yes including HN)


It would seem that they are "following the money".


Like dogs


Like capitalists...


Too late. They killed their own platform not providing good API services. When everybody was interested in connecting with Twitter they didn't provide options, now Twitter is falling and it would be very difficult to attract interest.


Is there a parameter to exclude tweets from bogus accounts? Twitter has clearly identified these and they've been disappearing by the millions, but will they still appear in the search archive?

It'd be nice to exclude those tweets, and given the very expensive API calls, only pay for non-human tweets when explicitly desired.


I would sooner pay $1 per Google query to expunge tweets from the results.


Only ~$1.00 per API call...


The default # tweets returned is 100, but the max is 500. That’s a $4 penalty for not using the correct API perameter.


Name really rolls off the tongue.


collect a bunch of results and print them out.

PFASEbook.


And, then Facebook could come out with their own framework named: Total Web-to-React-Redux framework...Or, TwtRR framework.




Join us for AI Startup School this June 16-17 in San Francisco!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: