LinkDigest · Schema

Digest

A post as LLM-ready text. chapters, chapter_source, category, music, image_count, images_read, coverage and key_point_support (added 2026-10-05) may be absent from links cached before then, as may links (added the same day).

Artificial IntelligenceSocial MediaData ExtractionSearchContent Aggregation

Properties

Name Type Description
platform string xiaohongshu, douyin, tiktok, youtube, x or web.
author string The author's display name.
title string The post's title. On Douyin its own title (item_title) when it has one, else the caption's first 120 characters; on YouTube the watch page's title when the page can be read, else Gemini's reading of t
posted_at string When it was posted (YYYY-MM-DD), when known.
caption string The post's own text.
transcript array What is said in the video, with times. Empty for image notes.
transcript_source string native_captions, asr or none.
ocr_text array Every distinct piece of text read from the images or frames.
on_screen array Video only: on-screen text with roughly when it appeared. Text shown for most of the video is in on_screen_persistent instead, once.
on_screen_persistent array Video only: the title card, watermark or slogan that stayed on screen, taken out of each on_screen entry.
frames array Video only (2026-11-07; absent on links cached before then): every sampled frame with its time, a short description and its text.
annotation object
images array Each post image described and read (image notes).
key_points array The substantive takeaways, in the post's language.
key_point_support array One quote per key point the model could quote, checked against the post by code. A point with no verified row has no evidence in the post's own words.
stats objectnull Engagement counts as the platform reported them at fetch time.
tags array The post's hashtags.
comments array The first comments the post's public share page served with it (Xiaohongshu: the top few), read with no second request and no login. Never the whole thread; coverage says how many of the post's commen
chapters array Chapters marked on the video by its author or the platform — never inferred. On a partial read, only those in the part read. Empty when the video has none.
chapter_source string Who marked the chapters: "author", "platform" (generated by Douyin), or empty.
outline array A timed outline of a long video: for a transcript of 8 minutes or more on a video with no marked chapters, 5 to 20 sections in order that cover everything read. Included in the read, no extra credit.
parts objectnull A multi-part Bilibili video's own parts (分P). Null for a single video and on links cached before 2026-11-10. Each part is its own read: send the same link with ?p=, one by one or as a batch.
category array The platform's own category for the post, e.g. 财经 › 投资理财 › 股票 on Douyin, Education on YouTube.
music string The soundtrack (title — artist) when it is not the poster's own voice; else empty.
links array The links the post points to — in its caption or description, WeChat's 阅读原文, QR codes in its images or frames, on-screen text and chapters — deduplicated, at most 20, and never opened: a link worth re
breakdown objectnull 爆款拆解 — how the post is built. Present only when breakdown=true (+1 credit). Quotes are checked against the post; engagement and tags come from the platform, never the model.
answer objectnull The caller's question answered from the post. Present only when ask was sent (+1 credit).
insight objectnull 消费者洞察 — what the post says about the people it speaks to. Present only when insight=true (+1 credit). Every item is a quote checked against the post and its top comments; nothing is invented.
translation objectnull Present only when translate_to was given (+1 credit). The originals are always kept.
raw_markdown string The whole digest as Markdown, including any breakdown and translation.
credits integer What this call cost. 0 when served from cache.
credits_left integernull Free accounts: credits left after this call (pack credits included). Absent on paying plans. In markdown, the X-Credits-Left header.
credits_note stringnull Free accounts with 3 credits or fewer left: the balance and the starter pack's pay link, in English.
offer_zh stringnull The same in Chinese, its link opening the ¥ Alipay page; for an agent to relay to its user.
cached boolean True when someone had already read this link — free.
degraded array Anything that did not read fully, said plainly.
duration_seconds numbernull Media length, when it has one.
read_seconds numbernull How much of the media was read.
partial objectnull Present only when partial_ok was set and the budget did not cover the whole video, or (since 2026-10-06) the whole image post: then image_count and images_read say how many images were taken.
depth string How the video was read: full (frames sampled densely), transcript (transcript-first: the whole transcript plus about a frame a minute, which coverage.not_captured says) or script (the speech only, no
auto_depth objectnull Present when the read was switched to transcript-first by the service (2026-11-05): a long video that a full read would not fit in the caller's credits, read whole at transcript depth instead of in pa
image_count integernull How many images the post has, on posts with images.
images_read integernull How many of them were taken and priced: every one up to 60; past that the first 60, and with partial_ok the first ones the budget covers; degraded says so. Taken is not read: a failed download or visi
coverage objectnull How much of the post this digest captured, computed in code. Null on links cached before 2026-10-05.
source_url string The link as resolved.
View JSON Schema on GitHub

JSON Schema

linkdigest-dev-digest-schema.json Raw ↑
{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "$id": "https://artifacts.apis.io/linkdigest-dev/json-schema/linkdigest-dev-digest-schema.json",
  "title": "Digest",
  "description": "A post as LLM-ready text. chapters, chapter_source, category, music, image_count, images_read, coverage and key_point_support (added 2026-10-05) may be absent from links cached before then, as may links (added the same day).",
  "x-generated": "2026-11-17",
  "x-method": "derived",
  "x-generator": "derive-json-schema.py",
  "x-source": "openapi/linkdigest-dev-openapi.json#/components/schemas/Digest",
  "type": "object",
  "properties": {
    "platform": {
      "type": "string",
      "description": "xiaohongshu, douyin, tiktok, youtube, x or web."
    },
    "author": {
      "type": "string",
      "description": "The author's display name."
    },
    "title": {
      "type": "string",
      "description": "The post's title. On Douyin its own title (item_title) when it has one, else the caption's first 120 characters; on YouTube the watch page's title when the page can be read, else Gemini's reading of the video."
    },
    "posted_at": {
      "type": "string",
      "description": "When it was posted (YYYY-MM-DD), when known."
    },
    "caption": {
      "type": "string",
      "description": "The post's own text."
    },
    "transcript": {
      "type": "array",
      "items": {
        "type": "object",
        "description": "One transcript segment.",
        "properties": {
          "t": {
            "type": "number",
            "description": "Start time in seconds."
          },
          "text": {
            "type": "string",
            "description": "What was said."
          },
          "approx": {
            "type": "boolean",
            "description": "True when t was estimated from the sentence's place inside a longer recogniser chunk."
          }
        }
      },
      "description": "What is said in the video, with times. Empty for image notes."
    },
    "transcript_source": {
      "type": "string",
      "description": "native_captions, asr or none."
    },
    "ocr_text": {
      "type": "array",
      "items": {
        "type": "string",
        "description": "A piece of on-screen text."
      },
      "description": "Every distinct piece of text read from the images or frames."
    },
    "on_screen": {
      "type": "array",
      "items": {
        "type": "object",
        "description": "Text on screen at a time.",
        "properties": {
          "t": {
            "type": "number",
            "description": "Approximate time in seconds."
          },
          "text": {
            "type": "string",
            "description": "The text on screen."
          }
        }
      },
      "description": "Video only: on-screen text with roughly when it appeared. Text shown for most of the video is in on_screen_persistent instead, once."
    },
    "on_screen_persistent": {
      "type": "array",
      "items": {
        "type": "string",
        "description": "A line of text shown for most of the video."
      },
      "description": "Video only: the title card, watermark or slogan that stayed on screen, taken out of each on_screen entry."
    },
    "frames": {
      "type": "array",
      "items": {
        "type": "object",
        "description": "One sampled frame.",
        "properties": {
          "t": {
            "type": "number",
            "description": "Seconds into the video the frame was taken."
          },
          "description": {
            "type": "string",
            "description": "What the frame shows, in a few words."
          },
          "ocr": {
            "type": "string",
            "description": "The text in that frame, verbatim."
          }
        }
      },
      "description": "Video only (2026-11-07; absent on links cached before then): every sampled frame with its time, a short description and its text."
    },
    "annotation": {
      "$ref": "#/$defs/Annotation"
    },
    "images": {
      "type": "array",
      "items": {
        "type": "object",
        "description": "One post image.",
        "properties": {
          "description": {
            "type": "string",
            "description": "What the image shows."
          },
          "ocr": {
            "type": "string",
            "description": "The text inside it, verbatim."
          }
        }
      },
      "description": "Each post image described and read (image notes)."
    },
    "key_points": {
      "type": "array",
      "items": {
        "type": "string",
        "description": "A takeaway."
      },
      "description": "The substantive takeaways, in the post's language."
    },
    "key_point_support": {
      "type": "array",
      "items": {
        "type": "object",
        "description": "The words behind one key point.",
        "properties": {
          "point_index": {
            "type": "integer",
            "description": "0-based index into key_points."
          },
          "kind": {
            "type": "string",
            "description": "Where the quote was found: transcript, on_screen, image, caption or chapter.",
            "enum": [
              "transcript",
              "on_screen",
              "image",
              "caption",
              "chapter"
            ]
          },
          "t": {
            "type": [
              "number",
              "null"
            ],
            "description": "Seconds into the video where the quote is (transcript, on-screen text, chapter)."
          },
          "image_index": {
            "type": [
              "integer",
              "null"
            ],
            "description": "0-based index into images, when the quote is in one image."
          },
          "quote": {
            "type": "string",
            "description": "A short phrase, verbatim from the post."
          },
          "verified": {
            "type": "boolean",
            "description": "True when the engine found the quote word for word in the post; false rows are kept."
          },
          "approx": {
            "type": "boolean",
            "description": "True when t is an estimate: a sentence placed inside a recogniser chunk, or a sampled frame's time."
          }
        }
      },
      "description": "One quote per key point the model could quote, checked against the post by code. A point with no verified row has no evidence in the post's own words."
    },
    "stats": {
      "type": [
        "object",
        "null"
      ],
      "description": "Engagement counts as the platform reported them at fetch time.",
      "properties": {
        "likes": {
          "type": [
            "integer",
            "null"
          ],
          "description": "Likes (点赞) the platform showed when the post was read."
        },
        "comments": {
          "type": [
            "integer",
            "null"
          ],
          "description": "Comments (评论)."
        },
        "shares": {
          "type": [
            "integer",
            "null"
          ],
          "description": "Shares (分享)."
        },
        "collects": {
          "type": [
            "integer",
            "null"
          ],
          "description": "Saves (收藏)."
        },
        "plays": {
          "type": [
            "integer",
            "null"
          ],
          "description": "Plays (播放), when the platform reports them."
        },
        "as_of": {
          "type": "string",
          "description": "The day the counts were read (YYYY-MM-DD) — a snapshot, not a live figure."
        }
      }
    },
    "tags": {
      "type": "array",
      "items": {
        "type": "string",
        "description": "A hashtag, without #."
      },
      "description": "The post's hashtags."
    },
    "comments": {
      "type": "array",
      "items": {
        "type": "object",
        "description": "One comment from the post's share page.",
        "properties": {
          "text": {
            "type": "string",
            "description": "The comment, verbatim."
          },
          "likes": {
            "type": "number",
            "description": "Likes on the comment, or null."
          },
          "replies": {
            "type": "number",
            "description": "Replies under it, or null."
          },
          "posted_at": {
            "type": "string",
            "description": "YYYY-MM-DD, or empty."
          },
          "reply_texts": {
            "type": "array",
            "items": {
              "type": "string",
              "description": "A reply, verbatim."
            },
            "description": "The first replies under it, text only."
          }
        }
      },
      "description": "The first comments the post's public share page served with it (Xiaohongshu: the top few), read with no second request and no login. Never the whole thread; coverage says how many of the post's comments these are. Empty on other platforms and on links cached before 2026-10-17."
    },
    "chapters": {
      "type": "array",
      "items": {
        "type": "object",
        "description": "One chapter.",
        "properties": {
          "t": {
            "type": "number",
            "description": "When the chapter starts, in seconds."
          },
          "title": {
            "type": "string",
            "description": "The chapter's title, as marked."
          },
          "summary": {
            "type": "string",
            "description": "Douyin's one-line summary of a chapter it generated, or empty."
          }
        }
      },
      "description": "Chapters marked on the video by its author or the platform — never inferred. On a partial read, only those in the part read. Empty when the video has none."
    },
    "chapter_source": {
      "type": "string",
      "description": "Who marked the chapters: \"author\", \"platform\" (generated by Douyin), or empty."
    },
    "outline": {
      "type": "array",
      "items": {
        "type": "object",
        "description": "One section.",
        "properties": {
          "start": {
            "type": "number",
            "description": "Where the section starts, in seconds: a transcript line's own time."
          },
          "end": {
            "type": "number",
            "description": "Where it ends: the next section's start, the last one the end of what was read."
          },
          "title": {
            "type": "string",
            "description": "What the section is about."
          },
          "summary": {
            "type": "array",
            "items": {
              "type": "string",
              "description": "One line."
            },
            "description": "One to three lines saying what is said in the section."
          },
          "quote": {
            "type": "string",
            "description": "A phrase from the section's own lines, kept only when it was found there word for word; else empty."
          },
          "generated": {
            "type": "string",
            "description": "\"model\" (the model's plan, checked by code) or \"fallback\" (equal stretches titled by their opening words)."
          }
        }
      },
      "description": "A timed outline of a long video: for a transcript of 8 minutes or more on a video with no marked chapters, 5 to 20 sections in order that cover everything read. Included in the read, no extra credit. Empty on shorter videos, on videos that have chapters, and on links cached before 2026-11-10."
    },
    "parts": {
      "type": [
        "object",
        "null"
      ],
      "description": "A multi-part Bilibili video's own parts (分P). Null for a single video and on links cached before 2026-11-10. Each part is its own read: send the same link with ?p=<number>, one by one or as a batch.",
      "properties": {
        "count": {
          "type": "integer",
          "description": "How many parts the video has."
        },
        "current": {
          "type": "integer",
          "description": "The part this digest is (the link's ?p=, 1 when it names none)."
        },
        "items": {
          "type": "array",
          "items": {
            "type": "object",
            "description": "One part.",
            "properties": {
              "p": {
                "type": "integer",
                "description": "The part number: read it with ?p=<number> on the same link."
              },
              "title": {
                "type": "string",
                "description": "The part's own title."
              },
              "seconds": {
                "type": "number",
                "description": "Its length in seconds."
              }
            }
          },
          "description": "The parts in order, at most 200."
        }
      }
    },
    "category": {
      "type": "array",
      "items": {
        "type": "string",
        "description": "One level, broadest first."
      },
      "description": "The platform's own category for the post, e.g. 财经 › 投资理财 › 股票 on Douyin, Education on YouTube."
    },
    "music": {
      "type": "string",
      "description": "The soundtrack (title — artist) when it is not the poster's own voice; else empty."
    },
    "links": {
      "type": "array",
      "items": {
        "type": "object",
        "description": "A link the post points to. Nothing behind it was read.",
        "properties": {
          "url": {
            "type": "string",
            "description": "The address as found, scheme and host lower-cased. wxp://, alipays:// and other app links are kept."
          },
          "found_in": {
            "type": "string",
            "description": "Where it was found: caption, description (YouTube), wechat_read_more (a WeChat article's 阅读原文), qr_code (decoded from an image or frame the read downloaded), on_screen (a well-formed URL in the on-screen or image text) or chapter.",
            "enum": [
              "caption",
              "description",
              "wechat_read_more",
              "qr_code",
              "on_screen",
              "chapter"
            ]
          },
          "label": {
            "type": "string",
            "description": "What it is when known (\"WeChat Pay\", \"QQ group\", \"阅读原文\"), else the words before it in the post, or empty."
          },
          "kind": {
            "type": "string",
            "description": "What opening it would do: web, payment (a WeChat Pay, Alipay or other payment code), group_invite (joins a WeChat, QQ, Telegram… group), app_deeplink (opens an app) or other (a contact card, an account to follow).",
            "enum": [
              "web",
              "payment",
              "group_invite",
              "app_deeplink",
              "other"
            ]
          }
        }
      },
      "description": "The links the post points to — in its caption or description, WeChat's 阅读原文, QR codes in its images or frames, on-screen text and chapters — deduplicated, at most 20, and never opened: a link worth reading is a separate digest. Payment codes and group invites are labelled by kind."
    },
    "breakdown": {
      "type": [
        "object",
        "null"
      ],
      "description": "爆款拆解 — how the post is built. Present only when breakdown=true (+1 credit). Quotes are checked against the post; engagement and tags come from the platform, never the model.",
      "properties": {
        "language": {
          "type": "string",
          "description": "The language the analysis is written in, e.g. zh-CN or en."
        },
        "format": {
          "type": "string",
          "description": "The kind of post, e.g. 口播 talking head, 图文笔记 image note, 教程 tutorial."
        },
        "hook": {
          "type": "object",
          "description": "How the post holds people in its first seconds (first image for an image note).",
          "properties": {
            "spoken": {
              "type": "string",
              "description": "The opening words, verbatim from the transcript, or empty."
            },
            "on_screen": {
              "type": "string",
              "description": "The opening on-screen or cover text, verbatim, or empty."
            },
            "visual": {
              "type": "string",
              "description": "What is shown at the start."
            },
            "why": {
              "type": "string",
              "description": "Why the opening holds attention."
            }
          }
        },
        "beats": {
          "type": "array",
          "items": {
            "type": "object",
            "description": "One step of the post's structure.",
            "properties": {
              "t": {
                "type": [
                  "number",
                  "null"
                ],
                "description": "Start time in seconds, approximate to a keyframe (video)."
              },
              "image": {
                "type": [
                  "integer",
                  "null"
                ],
                "description": "1-based image number (image notes)."
              },
              "label": {
                "type": "string",
                "description": "The step, e.g. 痛点 pain point, 方法 method, 行动号召 call to action."
              },
              "summary": {
                "type": "string",
                "description": "What happens in this step."
              },
              "quote": {
                "type": "string",
                "description": "A short line copied verbatim from the post, or empty."
              }
            }
          },
          "description": "The structure, 3 to 10 steps in order."
        },
        "title_formula": {
          "type": "string",
          "description": "The pattern behind the title, e.g. 数字 + 结果 + 人群."
        },
        "cover_text": {
          "type": "string",
          "description": "Cover or first-frame text, verbatim, or empty."
        },
        "cta": {
          "type": "string",
          "description": "The call to action, verbatim, or empty."
        },
        "audience": {
          "type": "string",
          "description": "Who the post is made for."
        },
        "tags": {
          "type": "array",
          "items": {
            "type": "string",
            "description": "A hashtag, without #."
          },
          "description": "The post's own hashtags, from the platform."
        },
        "engagement": {
          "type": [
            "object",
            "null"
          ],
          "description": "Engagement counts as the platform reported them at fetch time.",
          "properties": {
            "likes": {
              "type": [
                "integer",
                "null"
              ],
              "description": "Likes (点赞) the platform showed when the post was read."
            },
            "comments": {
              "type": [
                "integer",
                "null"
              ],
              "description": "Comments (评论)."
            },
            "shares": {
              "type": [
                "integer",
                "null"
              ],
              "description": "Shares (分享)."
            },
            "collects": {
              "type": [
                "integer",
                "null"
              ],
              "description": "Saves (收藏)."
            },
            "plays": {
              "type": [
                "integer",
                "null"
              ],
              "description": "Plays (播放), when the platform reports them."
            },
            "as_of": {
              "type": "string",
              "description": "The day the counts were read (YYYY-MM-DD) — a snapshot, not a live figure."
            }
          }
        },
        "template": {
          "type": "array",
          "items": {
            "type": "string",
            "description": "One step with [placeholders]."
          },
          "description": "4 to 8 steps to make your own version."
        },
        "remix_ideas": {
          "type": "array",
          "items": {
            "type": "string",
            "description": "An angle."
          },
          "description": "2 or 3 angles for a different version of the same idea."
        },
        "missing": {
          "type": "array",
          "items": {
            "type": "string",
            "description": "A missing signal."
          },
          "description": "What the analysis could not see (no speech, a partial read…)."
        }
      }
    },
    "answer": {
      "type": [
        "object",
        "null"
      ],
      "description": "The caller's question answered from the post. Present only when ask was sent (+1 credit).",
      "properties": {
        "question": {
          "type": "string",
          "description": "The question, as cleaned by the engine."
        },
        "answer": {
          "type": "string",
          "description": "The answer, from the post only; says the post does not say when it does not."
        },
        "found": {
          "type": "boolean",
          "description": "False when the post does not address the question — nothing was invented."
        },
        "support": {
          "type": "array",
          "items": {
            "type": "object",
            "description": "One quote behind the answer, checked by code.",
            "properties": {
              "kind": {
                "type": "string",
                "description": "Where the quote is: transcript, on_screen, image, caption or chapter."
              },
              "t": {
                "type": [
                  "number",
                  "null"
                ],
                "description": "Seconds into the video, when the quote has a time."
              },
              "image_index": {
                "type": [
                  "integer",
                  "null"
                ],
                "description": "0-based image index, when the quote is in one image."
              },
              "quote": {
                "type": "string",
                "description": "A short phrase the answer rests on, verbatim from the post."
              },
              "verified": {
                "type": "boolean",
                "description": "True when the engine found the quote word for word in the post."
              },
              "approx": {
                "type": "boolean",
                "description": "True when t is an estimate."
              }
            }
          },
          "description": "The quotes the answer rests on (1-3); empty when found is false."
        }
      }
    },
    "insight": {
      "type": [
        "object",
        "null"
      ],
      "description": "消费者洞察 — what the post says about the people it speaks to. Present only when insight=true (+1 credit). Every item is a quote checked against the post and its top comments; nothing is invented.",
      "properties": {
        "language": {
          "type": "string",
          "description": "The language the insight is written in, e.g. zh-CN or en."
        },
        "audience": {
          "type": "object",
          "description": "Who the post speaks to.",
          "properties": {
            "found": {
              "type": "boolean",
              "description": "False when the post and its comments do not support this section; items is then empty."
            },
            "items": {
              "type": "array",
              "items": {
                "type": "object",
                "description": "One finding and the quote behind it, checked by code.",
                "properties": {
                  "text": {
                    "type": "string",
                    "description": "The finding, in the insight's language."
                  },
                  "quote": {
                    "type": "string",
                    "description": "The words it rests on, verbatim from the post or a comment."
                  },
                  "kind": {
                    "type": "string",
                    "description": "Where the quote is: caption, transcript, on_screen, image, chapter or comment."
                  },
                  "verified": {
                    "type": "boolean",
                    "description": "True when the engine found the quote word for word in the post or its comments."
                  },
                  "t": {
                    "type": [
                      "number",
                      "null"
                    ],
                    "description": "Seconds into the video, when the quote has a time."
                  },
                  "image_index": {
                    "type": [
                      "integer",
                      "null"
                    ],
                    "description": "0-based image index, when the quote is in one image."
                  },
                  "approx": {
                    "type": "boolean",
                    "description": "True when t is an estimate."
                  }
                }
              },
              "description": "The findings; empty when found is false."
            }
          }
        },
        "claims": {
          "type": "object",
          "description": "The benefits and claims the post makes.",
          "properties": {
            "found": {
              "type": "boolean",
              "description": "False when the post and its comments do not support this section; items is then empty."
            },
            "items": {
              "type": "array",
              "items": {
                "type": "object",
                "description": "One finding and the quote behind it, checked by code.",
                "properties": {
                  "text": {
                    "type": "string",
                    "description": "The finding, in the insight's language."
                  },
                  "quote": {
                    "type": "string",
                    "description": "The words it rests on, verbatim from the post or a comment."
                  },
                  "kind": {
                    "type": "string",
                    "description": "Where the quote is: caption, transcript, on_screen, image, chapter or comment."
                  },
                  "verified": {
                    "type": "boolean",
                    "description": "True when the engine found the quote word for word in the post or its comments."
                  },
                  "t": {
                    "type": [
                      "number",
                      "null"
                    ],
                    "description": "Seconds into the video, when the quote has a time."
                  },
                  "image_index": {
                    "type": [
                      "integer",
                      "null"
                    ],
                    "description": "0-based image index, when the quote is in one image."
                  },
                  "approx": {
                    "type": "boolean",
                    "description": "True when t is an estimate."
                  }
                }
              },
              "description": "The findings; empty when found is false."
            }
          }
        },
        "objections": {
          "type": "object",
          "description": "Pain points, complaints and doubts — from the comments when they were read.",
          "properties": {
            "found": {
              "type": "boolean",
              "description": "False when the post and its comments do not support this section; items is then empty."
            },
            "items": {
              "type": "array",
              "items": {
                "type": "object",
                "description": "One finding and the quote behind it, checked by code.",
                "properties": {
                  "text": {
                    "type": "string",
                    "description": "The finding, in the insight's language."
                  },
                  "quote": {
                    "type": "string",
                    "description": "The words it rests on, verbatim from the post or a comment."
                  },
                  "kind": {
                    "type": "string",
                    "description": "Where the quote is: caption, transcript, on_screen, image, chapter or comment."
                  },
                  "verified": {
                    "type": "boolean",
                    "description": "True when the engine found the quote word for word in the post or its comments."
                  },
                  "t": {
                    "type": [
                      "number",
                      "null"
                    ],
                    "description": "Seconds into the video, when the quote has a time."
                  },
                  "image_index": {
                    "type": [
                      "integer",
                      "null"
                    ],
                    "description": "0-based image index, when the quote is in one image."
                  },
                  "approx": {
                    "type": "boolean",
                    "description": "True when t is an estimate."
                  }
                }
              },
              "description": "The findings; empty when found is false."
            }
          }
        },
        "purchase_intent": {
          "type": "object",
          "description": "Purchase-intent cues: 求链接, 多少钱, 哪里买 and the like.",
          "properties": {
            "found": {
              "type": "boolean",
              "description": "False when the post and its comments do not support this section; items is then empty."
            },
            "items": {
              "type": "array",
              "items": {
                "type": "object",
                "description": "One finding and the quote behind it, checked by code.",
                "properties": {
                  "text": {
                    "type": "string",
                    "description": "The finding, in the insight's language."
                  },
                  "quote": {
                    "type": "string",
                    "description": "The words it rests on, verbatim from the post or a comment."
                  },
                  "kind": {
                    "type": "string",
                    "description": "Where the quote is: caption, transcript, on_screen, image, chapter or comment."
                  },
                  "verified": {
                    "type": "boolean",
                    "description": "True when the engine found the quote word for word in the post or its comments."
                  },
                  "t": {
                    "type": [
                      "number",
                      "null"
                    ],
                    "description": "Seconds into the video, when the quote has a time."
                  },
                  "image_index": {
                    "type": [
                      "integer",
                      "null"
                    ],
                    "description": "0-based image index, when the quote is in one image."
                  },
                  "approx": {
                    "type": "boolean",
                    "description": "True when t is an estimate."
                  }
                }
              },
              "description": "The findings; empty when found is false."
            }
          }
        },
        "products": {
          "type": "object",
          "description": "Products, brands and prices named.",
          "properties": {
            "found": {
              "type": "boolean",
              "description": "False when the post and its comments do not support this section; items is then empty."
            },
            "items": {
              "type": "array",
              "items": {
                "type": "object",
                "description": "One finding and the quote behind it, checked by code.",
                "properties": {
                  "text": {
                    "type": "string",
                    "description": "The finding, in the insight's lan

# --- truncated at 32 KB (50 KB total) ---
# Full source: https://artifacts.apis.io/linkdigest-dev/json-schema/linkdigest-dev-digest-schema.json

Work with this as data

Every JSON Schema here is available over the APIs.io API and to AI agents over MCP.

MCP server

One button, every client — Claude, Cursor, VS Code and the rest.

https://apis.io/mcp

Tools for schemas

4 MCP tools reach this
  • find_json_schemasBrowse and filter every JSON Schema in the catalog.
  • apis_io_searchSTART HERE — APIs, providers and tags for one query, each with its total.
  • resolveTurn a domain, URL or GitHub org into the provider it belongs to.
  • find_cohortsEvery scored population of providers in the catalog.
All 92 tools →

Call it yourself

curl for this page
This JSON Schema
curl "https://apis.io/api/v1/json-schemas/linkdigest-dev-digest"
All schemas
curl "https://apis.io/api/v1/json-schemas?limit=25"

Calls without X-API-Key need no key. Calls with it need a paid plan — get a key.

Get an API key

Free tier, no form to fill in. Signing in shares your email address with us — we store it to create your key and to recognise you if you sign in with another provider. By signing up you are subscribing to the APIs.io and API Evangelist newsletters — you can unsubscribe from any issue. See our Privacy Policy and Terms.

A second provider on the same verified email joins the account you already have.