The Basics

  • How do I get started with Ingenia?

    Get in touch via the contact form and we'll send you your unique Ingenia API key, as well as your login details for the web interface.

  • How much does it cost?

    That depends on the way you will use it. Get in touch to find out all the details.

    We do feel confident enough you’ll love it to guarantee the following:

    • First month is free: so you can test the system and makes sure it’s not too good to be true
    • No charge for learning: Ingenia is tailored to your content, so it must learn from it. You can always send us as many "learning" items as you like, at no charge
  • How does it work?

    In short:

    • Ingenia can automatically categorize your textual content with your own tags
    • Your content should be organized in blocks of text that are somewhat self-contained and homogeneous in their content: we call them "items”
    • Based on its content, each item is associated with one or more “tags”, which represent what you think is important about that text, and that you define.
    • Similar tags can be grouped together in "tag sets"
    • For Ingenia to learn about you, we'll need you to send us a "training set" of items that have been already associated with their tags: we call this the "learning phase".
    • Once that's done, you're good to go: send Ingenia text that you want to tag, and you'll get the most relevant tags back in real time.
  • What is a good item?

    items don't need to be defined rigorously: for instance, most web pages are either an item (a newspaper article, a blog post, the description or the review of a product, an email, etc.) or a collection of them (your to-do list, a list of classifieds, the BBC homepage, your Facebook or Twitter stream, etc.). They can be as long or as short as you like: even just one word.

    You can define what constitutes an "item" as you like, and probably it'll be quite intuitive in your particular case, but just bear in mind that Ingenia will work best if they are like nuggets of content: somewhat self-contained and homogeneous. Get in touch if you’re in doubt

  • Can I host Ingenia myself?

    Yes. Ingenia is conceived to be used as an API, and we can host it for you on our servers, compliant with the best industry standards. However, we also offer a self-hosted solution, whereby we install Ingenia on your own servers and access it from there. This works well either if your data is very confidential, or if you want to ensure Ingenia scales in line with your requirements.

  • What exactly do you mean by tags?

    Tags are meaningful words or expressions that you want to associate to some or all your content items. They are the unsung hero of the internet: you find them in most content-rich web sites in some form, for instance:

    • Wikipedia: look for "categories" at the bottom of every page
    • News sites: each article is in a section, plus many sites, such as the New York Times or the Guardian, also have slightly more free-form tags, normally set by the journalist
    • Amazon: you have both "product categories", set by Amazon, and "customer tags" set by the users
    • even Youtube: for each video there are both "categories", where the video uploader can choose one of a few, and "tags", that are free form
    • ...and many more. In some sense, tags are similar to folders in your computer, with two big differences: (1) a file can only be in one folder, whereas an item can have several tags; and (2) we don't just associate an item with a tag, but we also give you the "strength" of its association, i.e., the confidence level.
    • So, tags are very flexible: they can relate to absolutely anything in the text, from the topic at any level of detail, to the type of content or its style, to any characteristic that is recurrent in your text: feel free to be creative. They work best if they are as specific and relevant to your content as possible. If in doubt, as usual, get in touch.
  • What is a tag set?

    Tag sets are thematically consistent groups of tags, such as, say, world countries, business sectors, product types, companies, concepts, topics, etc. You can keep it simple and have all your tags in your own big "tag set", or you can break them down into as many sets as you want: that will depend on your content and how you want to use Ingenia. You may begin with one, and perhaps break it down into more later on.

    Your tags, organised in your tag sets, represent your own "taxonomy" for your content. The more accurate, consistent and relevant your taxonomy, the better Ingenia will work for you.


  • What is the learning phase?

    In this phase, you’ll send us your “training set”: a sample of your items, each of them with tags that have already been added by you.

  • Why do I need to do a learning phase?

    As our service is tailored to you, we need to learn your own very specific preferences: which tags you want, and the unique way in which you associate your tags to your text. This changes hugely on a case by case, and the fact that we can learn from your own content is what makes the tags most relevant to you.

  • Can't you just apply some general tags?

    We believe that categorisation is useful if it's tailored to their users. There have been many attempts to create global categorisation systems, but they have never seemed as relevant as one that is built just for you.

    However, we are planning to build several tag sets that we believe are probably relevant to many people, such as "countries in the world" or "olympic sports" or, like in the demo, "food": you will be able to import them and to apply them to your content; you will also be able to import them and change them to make them yours. Watch this space - or get in touch if you want to learn more.

  • How many items should be in the training set?

    This can vary, depending on how well organized and tagged your content is; as a ballpark figure, if you want to use, say, N tags, you should ensure that in the training set each tag is used for at least 10 items. Obviously if you have more, send them through, as Ingenia will learn about you more quickly and accurately.

  • What if my items are tagged, but not very well?

    No worries, that’s almost always the case. With Ingenia you can improve this via iterating: send us what you have, via our web interface you’ll be able to spot where the main problems lie, get our recommendations, and amend your tags until you’re satisfied with it. It will be a very interesting process, and perhaps even fun. (At least we think so.)

  • What if my content is not already tagged?

    A person familiar with the content and the business should spend some time thinking about what tags are most relevant to it, and then tag a sample of items, sized as described above. Creating this initial training set shouldn’t take this person more than 10-20 seconds per item, or 20-30 minutes for 10 tags and 100 items. We’re very happy to help with this.

  • How do I need to send the training set?

    We recommend you send the training set in one block in the beginning: that will simplify your experience; however, you will always be able to send as many learning requests as you want at any point: Ingenia is continuously learning about you, and the more items you send, the better it will understand how you want to tag your content.

  • What if I have new tags?

    You can introduce new tags at any point: just send us a few items tagged with them. In theory one would suffice, but we do recommend at least 10 per tag, ideally more. If that tag is a new one but you think it applies to old items too, feel free to include these items in the training set, but if you do so make sure you include all the tags that you want to associate to each of them.

  • What if I want to change the way I tag my items?

    That happens often: just resend us the items, with their new tags, and we'll automatically take them into account as we continuously adapt to your content.


  • How long before I can use the service to tag my content?

    Once you have sent us your training set, after a few hours at most Ingenia will be ready to tag your new content. You can check the status of your data via the /status API request, or via the web interface.

  • What are these confidence levels?

    They represent the strength of the association between the item and the tag. In other words, the classification tells you not only that your item is associated with, say, the tag "Technology", but also "how much" it is associated with it. This can be used in many ways, for instance, in order to prioritise items in a search, or for a recommendation engine.

  • What response time should I expect?

    The response for a classification aims to be real time, i.e., within the current call. In the rare cases in which that doesn't happen, you’ll receive the ID of the item with a classification status "pending". In this case, you can retrieve your tags either by sending us the same text again (even more than once, we won’t create duplicates). If there is a problem with the call, we'll indicate it in the "status" field.

  • How can I keep an eye on the way the API is working for me?

    You can monitor the usage of the API via the web interface, where you can:

    • Have an overview of your tags, how often they are used, how well Ingenia understands them
    • Manage them: delete or rename them
    • Be alerted if Ingenia has any action to recommend, based on our own statistical analysis

What's next

  • Are you introducing any new features?

    Plenty: we're just started. We can't share all of them quite yet, but here are some:

    • Rules on allocating tags, for instance, whether or not each document must have at least, or at most, one
    • Manage different sections of your item (say, title and content) differently (say, paying extra attention to words in the title)
    • Fully fledged web interface, to manage your tags, visualise your data and spot opportunities for improvement
    • Feature-rich web crawler, to categorise directly your website
    • Shareable tag sets: tag sets that are relevant to several common types of content, that you can import and use for your content without having to go through the learning phase

    And if you have any idea or suggestion, please let us know via the contact form, we'll do our best to integrate it in the API as fast as we can.

Data treatment and Your Account

  • What happens to my data?

    Your data is yours: if it’s not highly confidential, we recommend leaving it with us, as we’ll soon be able to offer you new services that make use of it, such as advanced visualisation and analytics.

    If you prefer us not to keep it, we have an on-premise solution, whereby you'd host your Ingenia instance yourself, so that your content never gets out of your servers.

  • Closing Your Account

    We’re confident you’ll love Ingenia, but in case you want to close your account, we’ll send you all the associations between tags and items that we created for you in JSON, and then we’ll permanently delete all your data.