Annual subscription contracts offer our best prices with easy monthly billing for a committed number of monthly transaction credits. Any monthly overages you use are charged at the same per-transaction rate as your monthly subscription price.
|Number of transaction credits||Monthly subscription price|
|1 thousand transactions per day||FREE||Get Started|
|90 thousand transactions / month||$250/mo||Choose Plan|
|300 thousand transactions / month||$750/mo||Choose Plan|
|3 million transactions / month||$1750/mo||Choose Plan|
|High volume media intelligence, 1 Billion+ transactions / month||Contact Sales||Contact Sales|
Every account gets the Free plan. Even if you’re just trying us out, you get email support to help you succeed.
|Performance & Support Features||FREE plan$0||Pro plan$595||Enterprise plan$1995|
2 million tx/mo1 concurrent connection
10 million tx/mo5 concurrent connections + Combined Call
100 million tx/mo50 concurrent connections + Combined Call
Contact us for higher throughput rates
Email supportWe answer everyone, though higher level plans get faster response times
Phone supportCall us during business hours, or leave us a voicemail, and we'll help you with any issue you might have.
SLA uptime guaranteeUptime within our commitments or we’ll credit your account
Monthly subscription contracts give you the flexibility of our shortest commitment combined with pay-for-what-you-use pricing. When your application reaches volume scale, you'll automatically get volume pricing. No need to contact us. All month-to-month subscription plans come with our Free performance and support package.
|Number of transactions per month||Cost per transaction|
|First 250 thousand transactions / month||$0.0035 per transaction|
|Next 750 thousand transactions / month||$0.0015 per transaction|
|Next 1 million transactions / month||$0.00075 per transaction|
|Next 3 million transactions / month||$0.00050 per transaction|
|All transactions over 5 million per month||$0.00035 per transaction|
A transaction is a volume pricing unit based on the number of API functions you execute on a text and the computational intensity of the API function itself. Most AlchemyLanguage API functions use one transaction unit when they are invoked to analyze a text, with exceptions being some low-level utility calls like constituent parsers that we make available on request. For example, if a text is first classified with the taxonomy API, and then analyzed with both the keywords and sentiment APIs, then that counts as three transactions.
The texts you analyze are size limited, but the vast majority of documents can be processed with a single transaction unit. URLs have a maximum size allowed of 600 kilobytes. A raw text document is limited in size to 50 kilobytes.
To put these size limits in perspective, Google reported the median size of an HTML page is 176 kilobytes, with 90% of pages from top sites coming in at under 597 kilobytes per page. For sizing raw text, a character takes up a byte. So a kilobyte is a thousand characters, or roughly 100 words. At about 2 kilobytes per typewritten page, the 50 kilobyte size limit on text would process the equivalent of 25 pages.
On the rare occasions you need to analyze bigger documents you can chunk them into smaller sizes before processing them with AlchemyLanguage. We intend to give you other options in the future.
We charge your credit card each month for the committed number of transaction credits per month plus any transaction overages you had the previous month. Transaction credits reset to zero on the first of every month. Overages are charged at the same per-transaction rate as your subscription, or roughly (your subscription price) / (number of transaction credits). For example, the per-transaction overage cost for a 3 million transaction-per-month subscription plan would be $1750 / 3,000,000 = $0.00058 per transaction.
You can increase your committed transaction number at any time and charges will adjust with the next monthly bill. Your account will automatically renew at the end of the year unless you change your subscription. Please note you cannot lower your committed transaction volume or cancel your annual subscription until the contract ends.
We offer a 10% discount on your base subscription cost if you pay for a full year’s base subscription up-front. We’ll still bill your credit card monthly for any overages you used in the previous month at the pre-discounted rate.
Using the month-to-month subscription schedule, we charge your credit card each month for the number of transactions you used in the previous month. For example, if I used 2 million transactions last month, I would be billed (250,000*$0.0035) + (750,000*$0.0015) + (1,000,000*$0.00075) = $2,750.
You can cancel the service any time. Your account will automatically renew each month unless you change your subscription. This plan might fit your needs while you are ramping up your AlchemyAPI deployment, processing a large backlog of texts, or if the monthly volume of texts you process fluctuates from zero up to hundreds of thousands and beyond.
AlchemyAPI is the only NLP cloud service that can support the throughput demands of applications in high-volume media monitoring, financial trading and the real-time bidding systems for online display advertising. The performance plan you need depends on your peak throughput demands. Our high-speed cloud infrastructure supports parallel computing where multiple transactions can be executed concurrently across numerous cpu cores. In addition, our Combined Call feature in the Pro and Enterprise plans lets you increase your results throughput when executing multiple NLP calls against the same web page. Let’s look at throughput a little deeper.
AlchemyAPI does all the heavy lifting of acquiring clean text from web pages to make them ready for processing. The total time required to download a typical blog or news site URL, crawl, extract and clean its text, analyze it with a single API call, and return results usually range from 300-900 milliseconds (tweets and raw texts are processed faster).
In our Free plan, this means a single processor core can conservatively average a transaction per second, or with roughly 2.6 million seconds in a month, 2 million transactions per month on a steady flow of URLs. Our Pro plan offers 5 concurrent connections and can process 10 million URLs each month with a single NLP call. Scaling up to higher monthly averages is relatively simple, with our Enterprise plan offering 50 concurrent processing connections and a throughput of over 100 million URLs per month. Even greater concurrency is available with special pricing options, so please contact us.
Many customers execute several NLP functions on a single URL. Rather than repeating the text acquisition cycles for each NLP function, which can eat up 25%-50% of the total processing time, our Combined Call lets you increase your throughput by performing these steps once per URL and then processing the text with all the calls you choose. For example, if you process every web page with five APIs - Taxonomy, Concepts, Entities, Keywords and Sentiment - the Combine Call would eliminate four of the usual text acquisition cycles. You’ll find the Combined Call turned on for Pro and Enterprise plans.
Of course, it’s usually the peak load you design for, and the simple rule of thumb that one processor core has a throughput of 1 URL per second for a single NLP call should help you decide which performance level you need: The Free Plan limits you to 1 processor core, so for more than 2 million transactions/mo upgrade to our Pro plan. For more than 10 million transactions/mo go to the Enterprise plan.
AlchemyAPI is unique among NLP providers in that nearly all of our customer’s billions of monthly transactions use the same API calls to correctly understand content across wildly disparate industries and knowledge domains. This is made possible by our deep-learning AI breakthroughs we have been using to continually train our algorithms since 2009 on truly massive web data. But some applications benefit when you can classify text into your company’s specific product categories, or when you can extract named entities completely unique to your domain. Our Enterprise Plan gives you access to our services team where we make available tools and paid consulting services to tune the Taxonomy and Named Entities APIs to your needs.
Yes. You can download a free key to the AlchemyLanguage service here and get started right now. You will have the same access to all standard AlchemyLanguage functions as the paid plan, but your transaction volume is capped at 1000 transactions/day and you cannot execute functions in parallel. When you decide to increase your transactions, processing speed or customize the functions, you can contact our sales team and upgrade to a subscription plan that suits you.
You can upgrade your annual subscription plan any time by contacting our sales team. Annual contracts have a committed number of monthly transaction credits. Any monthly overages you use are charged at the same per-transaction rate as your monthly commitment. Higher commitment levels give you cheaper costs per transaction, so it’s wise to sign up for the subscription level that fits you best. It’s fine to upgrade to a higher monthly transaction commitment part way through your annual contract, but note you cannot downgrade or cancel your annual contract during its active term.
Month-to-month contracts offer pay-for-what-you-use flexibility at a price premium to the annual contracts. You automatically get volume pricing as your application reaches volume scale. No need to contact sales.
We encourage you to compare your existing plan to what is currently available, but you can continue to use your existing subscription without changes to your pricing or available feature set while your contract is active. Changes can take place at your contract end date, or if you choose to upgrade your plan by adding more transactions, increase your performance or wish to customize your functions. Migrating to one of AlchemyAPIs standard pricing plans is easy, details are listed above.
You can get in touch with us by visiting our contact page.