BBC News Labs Documentation

Documentation for BBC News Labs APIs & #newsHACK events

Language Tech #newsHACK

BBC Connected Studio and News Labs have partnered with the World Service to research and develop tools in language technology, and we are inviting academic institutions and professionals involved in this specialism to take part.

We held a Language Technology #newsHACK in December 2014, and a similar event is now being held in London on 15-16 March 2016.

We're looking for teams to participate who will bring one or more instances of a 'Language Technology' to use in prototype ideas that they will work on during the two-day hack event. We are making some space for interested individuals (eg multilingual journalists, or UX designers) who are not bringing their own language tech, and we'll be integrating them into teams that are, where appropriate.

We are looking for concepts utilising:

  1. Speech to text (ideally in a non-English language)
  2. Text to Speech (ideally in a non-English language)
  3. Auto Translation between languages
  4. Speech Synthesis
  5. Entity Extraction and NLP in non-English
  6. Multilingual graphics
  7. Surprise us! (another tech involved in multilingual machine-driven systems)

BBC Developer Resources

Content Store API

The BBC Content Store API (aka CANDY) provides an interface to BBC News content with structured metadata.

CANDY stands for "CPS and Dynamic", it integrates the Content Production System (CPS) with the Dynamic Semantic Publishing system (DSP).

The Content API is a production platform, used internally by the BBC on the BBC website and in BBC mobile apps.

In addition to English, the Content API includes articles written in several languages.

Check out the documentation for the Content Store API

API key available for the event duration only: Bd1QARJ8BD2rAfE9i9igC8IlDR6ywXVG

Juicer API

The Juicer is a tool for automatic entity extraction and archiving from BBC News Labs.

You can use it fetch the latest news as well running queries for news coverage from the BBC and other news organisations.

Due to rights restrictions we are only able to link to third-party content and cannot republish anything on our services.

Check out the documentation for the Juicer API

API key available for the event duration only: Yk3KFzyVliTamletdhg7FOAqf26vwGeL

Example uses of the Juicer

Here are a few examples of projects and prototypes News Labs has created with Juicer:

Mango API

This API has been built by BBC R&D as part of the ABC-IP project. It extracts topics from text, using an algorithm similar to the one we used for automatically tagging the World Service archive. The topics extracted are DBpedia identifiers.

For the duration of the event, we provide a version of the Mango API supporting the following languages:

  • Arabic
  • Chinese
  • English
  • Spanish
  • Portuguese
  • Russian

API URL and documentation: http://mango.newshack.newslabs.co/

Third-party Resources


Questions and Feedback

All comments are moderated for approval.