Data source types

A data source brings articles into Nuvelics. Everything downstream, from events to pitches, starts from what your sources collect. The Add Data Source form offers ten types. This page says what each collects, which credential it needs, and what to ask your provider for. How to add, run and read a source is in Data Sources, a first walk-through with a feed is Add a data source, and how to store a credential is in Credentials.

The types at a glanceCopy Anchor Link

Every type also takes an optional Category and Location, which tell Nuvelics what the source mostly covers and help it route that source's articles to the right desks.

About credentialsCopy Anchor Link

• A credential is stored once under Credentials and attached to any number of sources. It must be made from the type that matches the source: an AP Media source takes only an AP Media credential, so a key for one service cannot be used against another.
• Creating a credential needs the Owner or Admin role. A Member who adds a source that needs one asks an administrator to create it. For the vendor types the form lists only the credentials that fit.
• The form requires a credential for Datafactory, DigiChief, AP Media, Reuters and AFP, and for X unless your organization uses Nuvelics's own X access. For RSS Feed and xAI Search it is behind a switch, and Council Meetings takes none.
• A credential's Health shows whether the last use worked. If a source shows Errors or Stale, open its credential first.
• To ask your provider for: an account with API access to the product, and the values listed under each type. Subscriptions, plans and quotas are between you and the provider; Nuvelics reads only what your account is entitled to.

RSS FeedCopy Anchor Link

Collects the articles of any feed you can open in a browser. Give it the feed's address (https only, with no sign-in details in the address). Turn on Enrich with full article body when the feed carries only summaries and you want the full text of each article as well.

A feed that sits behind a login needs a credential: switch on This endpoint requires authentication and choose one of the generic types, API Key, HTTP Basic or HTTP Bearer. Attaching one needs the permission to manage credentials.

X · Follows and X · ThreadCopy Anchor Link

X · Follows watches a list of X accounts, up to 100 usernames without the @, and collects the posts each originates, up to a number you set per account (20 by default). X · Thread takes one numeric tweet ID and collects the author's whole thread, the post and the author's own follow-up posts, wherever in the thread the ID points. For both, Follow links also reads the article a post points at, so a headline-and-link post is collected with the story's full text.

Both need an X credential, a bearer token from your X developer account. If your organization has been set up to use Nuvelics's own X access, the credential is optional and the form shows Use your own X API key instead; ask Nuvelics if you want this. These are not the same as the X account you publish from, which is connected under Social Media.

xAI SearchCopy Anchor Link

Runs a live search for a Search location, free text such as a city, with no geocoding. Choose the search mode: X only, Web or X + Web. It keeps up to a set number of stories (20 by default) and drops any older than the Recency (hours) you set (72 by default).

It needs an xAI credential, an API key that begins xai-, which is optional: with Use your own xAI API key off, Nuvelics searches with its own. Each xAI Search source makes its own paid searches, so two sources for the same place cost twice.

DatafactoryCopy Anchor Link

Collects match results from a football Competition, given as its slug, such as premierleague, laliga or seriea. It reports finished matches only, looking back a number of hours (36 by default) and returning up to a number of articles (20 by default). Only the Results profile exists; upcoming and live matches are not collected.

It needs a Datafactory credential, the API key from your Datafactory account.

DigiChiefCopy Anchor Link

Collects market data or sports data as articles. Choose a Profile:

• Top movers (under Financial) reports the most active symbols (10 by default).
• Watchlist (under Financial) reports every symbol you list, up to 50.
• Results (under Sports) reports finished games and Forward looking the schedule, for one League: NBA, MLB, NFL, MLS, Champions League or Premier League (20 games by default).

It needs a DigiChief credential with the Customer ID and API key DigiChief issued to you. The Customer ID is kept in the credential, not on the source.

AP MediaCopy Anchor Link

Collects the wire articles of one AP product (for example a general news product), up to a set number per run (25 by default). The form lists the products your credential is entitled to; choose one, and create a separate source for each product you want.

It needs an AP Media credential, the API key from your AP Media API account.

Good to know:

• Check with AP whether your product is an ongoing subscription: some products are provisioned for a limited time, and the product list shows each plan's duration. When a plan lapses, the source fails with the same error as a wrong product.
• The entitlement check runs only in the web form. A source saved through the API or the CLI with a product your account does not have is accepted, then fails when it harvests.
• AP's usage terms travel with each story and its pictures and are shown to your editors, but Nuvelics does not restrict where AP-sourced text can go. Decide that against your AP agreement.
• When AP withdraws a story, Nuvelics marks the stored article as recalled and does not use it for new pitches. A cluster that already holds the story can still be pitched, a recall is missed if the source was down too long, and nothing already pushed to your newsroom system or CMS is retracted.

ReutersCopy Anchor Link

Collects wire articles from Reuters' Content API. Narrow them by Categories, Regions, Languages, Sources and Channels, which are loaded from your Reuters account. An empty filter means no restriction, not "match nothing". Harvesting is text only for now, so Media types is fixed.

It needs a Reuters credential: the Client ID and Client secret from your Reuters Content API account.

Good to know: Nuvelics does not show your Reuters points balance, and cannot tell a source that fails because points ran out apart from other Reuters errors. Watch the source's status.

AFPCopy Anchor Link

Collects wire content from the AFP API. Choose a Content type, Text, Picture or Video; one source covers one type, so make a separate one for each. Channels narrows it to channels you pick, loaded from your AFP account; leave it empty to take every channel of that type in your account.

It needs an AFP credential: the Client ID and Client secret AFP issued for your API access, and the Username and Password of an AFP account. AFP recommends a dedicated service account for an automated feed, and it needs only the user role, not manager.

Good to know: Picture sources can be created now, and the source shows Picture harvesting not ready. They begin collecting once picture harvesting is released; Text and Video collect today.

Council MeetingsCopy Anchor Link

Collects the recent meetings of a city council from its meeting portal: agendas, minutes and their supporting documents. Give it the portal's address, the one you open in a browser, such as a Legistar or CivicClerk site. Nuvelics works out the platform from the address; choose Legistar or CivicClerk under Platform only if the portal sits behind a custom domain it cannot recognise. It returns a set number of recent meetings (10 by default).

It needs no credential: both platforms are public.

Manage it another wayCopy Anchor Link

• CLI: nuvelics datasource and nuvelics credential.
• API: Data sources and Credentials.
• AI assistants: the data source tools.