Videos, podcasts, events and research for app developers

This page is for teams building their own Arcel Konnect screens on the Daily Dose API: the videos and podcasts of Discover (to embed in your own player), saving, liking, sharing and commenting, the reader's collection, events, research papers, publisher logos and in-app feedback. Every operation named here links to its entry in the reference, with its inputs, replies, errors and examples.

Listing and opening videos and podcasts

The API gives you the items and the fields to play them; your app draws the screens and embeds the player.

• listExplore with kind set to video or podcast lists the items a page at a time (sort latest or popular, the filters of the reference, and cursor for the next page). getExplore lists the kinds Discover offers.
• getExploreItem opens one item.
• listUpNext gives the suggestions to show after the item, with their reasons.

Each video or podcast from YouTube carries details.player:

Field	Use
videoId	The 11-character id, for YouTube's native player libraries on iOS and Android
embedUrl	https://www.youtube-nocookie.com/embed/<id> (YouTube's privacy-enhanced domain); add playsinline=1 and your own player parameters
watchUrl	https://www.youtube.com/watch?v=<id>: open it in the YouTube app or the browser; always offer it
embeddable	true, false (the owner does not allow embedding: open watchUrl instead) or null (not checked yet: try the player, and fall back to watchUrl)

The item's image is the video's thumbnail with its credit, and details.channel names the channel (name, handle, url).

// List the latest videos and choose how each one plays.
const res = await fetch(`${API}/v1/dailydose/explore/video?sort=latest`, {
  headers: { Authorization: `Bearer ${sessionToken}` },
});
const { items } = await res.json();
for (const item of items) {
  const player = item.details.player;
  const play = player?.embeddable === false ? player.watchUrl : player?.embedUrl;
  console.log(item.title, play);
}

YouTube's rules

Follow YouTube's API Services Terms and its required minimum functionality, with the IFrame Player API or YouTube's native player libraries:

• Show YouTube's player whole and unobstructed, at least 200 × 200 px; never cover its controls or branding.
• Never play the audio alone: a YouTube podcast plays in the video player too.
• Never download or cache the video; use embedUrl or watchUrl exactly as the API gives them.
• When embeddable is false, open watchUrl instead of the player.

Curated items: fields that may be empty

ARCEL's staff add many videos and podcasts by pasting their links, and the API reads their title, channel and thumbnail from YouTube's public oEmbed. Until the YouTube Data API is connected, such an item has:

• publishedAt: null (unknown, never guessed). Show details.curated.at instead, for example "Added 6 Oct 2026".
• details.durationSeconds: null. Hide the duration.
• details.channel.id and details.channel.avatar: null. Show the channel's name, and its initial in place of the picture.

Up next and playback events

Report playback with captureEvents, with the item's exploreItemId: Explore Play when it starts, Explore Progress at 25, 50, 75 and 95 %, Explore Complete at the end, and Up Next Open when a reader opens a suggestion (with its position and reason). The counts on the cards and the Popular order are computed from them.

Actions on an item

Every item lists the buttons to draw in actions; draw only those. The API refuses an action the kind does not offer with 422 ACTION_NOT_AVAILABLE.

Kind	actions	Counts to show
event	save, share	none
podcast, video	save, share, comment, like	counts.likes, counts.comments
research	read, save, share	counts.reads

• Save: saveExploreItem and unsaveExploreItem; saved says whether the reader saved the item. Guests may save; their saves move to their account when they sign in.
• Like: likeExploreItem and unlikeExploreItem, for videos and podcasts only: a like on an event or a research paper answers 422 ACTION_NOT_AVAILABLE.
• Share: getExploreItemShareLink gives the link to share; report Share with captureEvents.
• Read: call recordExploreRead when a research paper's detail opens (only for research: a read counts once per reader).

Show your app's change at once and keep the reply's state and counts; on an error, put the button back as it was.

Comments on videos and podcasts

Comments on a video or a podcast work as comments on a story, with their own operations: listExploreItemComments, createExploreItemComment (a reply names its parentId), listExploreCommentReplies, updateExploreComment, deleteExploreComment and restoreExploreComment within the undo time, likeExploreComment, unlikeExploreComment and reportExploreComment. A guest is asked to sign in first (403 SIGN_IN_REQUIRED); a comment the community rules hold says it waits for a review (state: held).

The reader's collection

What a reader saved, liked and wrote, across stories and Discover, is private to them. getCollectionSummary gives the counts of each section by kind, and listCollection lists one section (saved, liked or comments) and kind, newest first, a page at a time. Each entry holds a story or an item; an entry whose story or item is no longer available (or an event that has ended) has available: false and keeps its title in unavailable.

Events

List events with listExplore and kind=event (upcoming, in date order) and open one with getExploreItem. The detail's sections follow the event detail design: the summary, "Event Details", "What you'll experience", "Pillars covered", "Technologies covered", "What you'll take away" and "Who should attend".

Field	Show it as	When it is empty
summary	The description under the title	The organiser's own description; when the organiser gives none, an AI-assisted one from the organiser's page (aiSections.fields names summary); null when neither exists
details.startsAt, details.endsAt, details.timeZone	Date and time in the event's own zone	timeZone (IANA, for example America/New_York) is the organiser's, else the event city's when the start's offset agrees with it; absent when unknown: show startsAt with its own offset
details.venue, details.address	The venue and its full address	Absent for online events or when the organiser gives none
details.city (code, label), location.cities	The place; filter with city	city.code is an IATA city code, null when the city has none (the name stays in label); location.cities is then []
details.organiser	"by …"	Absent only when neither the page nor ARCEL's source review names one
image, details.images	The hero and the carousel	details.images holds the extra images only (empty when the organiser gives one)
details.highlights	"What you'll experience"	The organiser's list, else AI-assisted; absent when the page supports none
details.pillarHighlights (pillar, name, description)	"Pillars covered": name as the heading	Only the event's own pillars; absent when none is supported by the page
details.technologies	"Technologies covered" chips	Absent when the page names none
details.takeaways	"What you'll take away"	Absent when the page supports none
details.audience	"Who should attend"	Absent when the page says nothing about it
details.attendeeCount	Attendees	Only a count the source publishes as attendees or registrations (never capacity, saves or views); no current source publishes one, so it is absent
details.expectedAttendeeCount	Expected attendance	Only when the organiser's page states the attendance it expects for this edition
aiSections (label, generatedAt, fields)	Show label with the AI-assisted fields	Absent when nothing on the card is AI-assisted
publishedAt	Not needed: date an event by details.startsAt	Usually null (organisers rarely declare when they published the page; never guessed)

The AI-assisted fields are written from the organiser's own event page and checked against it as story summaries are; a field the page does not support is left out rather than filled.

Publisher logos

Story cards in getBriefing and getStory carry publishers[].logo ({ url, credit }): a short-lived address of the logo ARCEL stored from the publisher's own site (a square icon where the publisher declares one). It is null for a publisher whose site declares no usable raster logo or whose logo host refuses crawlers; show the publisher's first letter then. Do not cache the address beyond its expiry; ask again.

Research papers

Research items come from reviewed publishers and research indexes; the API stores their metadata and, where the licence allows, their abstract, never the full text. List them with listExplore and kind=research (sort latest or popular, and the paperType, pillars, country and publishedSince filters), and open one with getExploreItem.

Field	Show it as	When it is empty
shortTitle	The card's title (at most 70 characters, only words of the paper's own title, AI-assisted)	Absent until one is made: show title
title, publisher, topics	The detail's head (the full title)	Always given
image	The card and detail picture, with image.credit	kind: illustration is ARCEL's shared AI illustration for the paper's pillar and type (not a picture of the paper), kind: generated an AI image made for this paper; null only before the illustration set exists (show your own)
details.paperTypeName	The paper's type (Research Paper, Market Report…), the display name of details.paperType. This is the readable label you asked for as details.paperTypeLabel: paperTypeLabel stays the publisher's own wording ("journal-article"), not for display	Absent when the paper has no controlled type
details.authors, publishedAt	The author block, "Published on …"	Authors absent when the metadata names none; publishedAt null when the source declares no date (never guessed)
details.doi	The DOI	Absent when the paper has none (reports, most library pages)
summary	The one-line "See how…"	Null for metadata-only papers (most of them): their licence allows no text from the paper
aiSummary.paperSummary with aiSummary.label	"Paper summary", with the label beside it	Null when there is no licensed source text to summarise, or before the summary is made: show details.abstract (when given), else nothing
aiSummary.whyItMatters	"Why it matters?"	Hide the section when aiSummary is null
details.abstract	The abstract	Given only when the work's licence (or a recorded permission) allows it
details.fullPaperUrl	"View full paper": open it outside your app (the publisher's page or the open-access copy), and report Research Paper Opened with captureEvents	Always given
counts.reads	"631 reads" on the card	0

A paper has readable text in at least one of aiSummary.paperSummary, details.abstract and summary whenever its licence lets ARCEL keep any of the paper's own text; a metadata-only paper has none of them, and nothing is generated from its title alone.

When the detail opens, call recordExploreRead.

In-app feedback

Readers and guests can send feedback from any screen:

1. getFeedbackOptions gives the topics ("Feedback about", in their order), the types (Problem, Suggestion, Other), the longest text and the image limits (how many, how large, which types). Start on the topic of the screen the sheet was opened from.
2. For each image: createFeedbackUpload with its type and size, then a PUT of the file to uploadUrl with the same Content-Type. Convert HEIC photos to JPEG first. The server removes the images' metadata (location, camera, dates) before staff see them.
3. submitFeedback with the topic, the type, the text, the uploads' keys and the app's context (appVersion, platform, and when you know them osVersion, device, screen, locale, timeZone), with an Idempotency-Key header (a new UUID per feedback, reused on a retry). Then thank the reader: there is no reply channel.

Keep the reader's topic, type and text on the device until the feedback is sent, so closing the sheet keeps the draft; never keep the images.

Your data

A signed-in reader can download everything ARCEL holds about them with exportAccountData, a JSON file, from your privacy settings.