Substack by Category rank_change via get_top_trends

Pass sort rank_change plus an official Substack category on get_top_trends to list climbers, then compare 3d fills against empty 1d notes and the overall Substack board from one account.

Live data as of 2026-08-30

Substack by Category rank_change on Trends MCP is get_top_trends with type Substack by Category, an official category name, sort rank_change, and window 3d. From the August 30, 2026 pull, in-sync boards stamped as_of 2026-08-30T08:01:36 UTC. Technology 3d used baseline 2026-08-27T02:01:32 UTC and returned count 14. LearnAIWithMe sat at rank 8, prev_rank 80, rank_change 72. Windows 1d and 7d returned count 0 with a complete-turnover note even though baseline_ts was present. Windows 14d and 30d set baseline_ts to null. Omitting window defaults to 30d and returns that empty body. type Substack filled 21 climbers on window 1d. Free tier access is 100 requests per month.

Technology 3d closed at 14 climbers

From the Trends MCP scan of August 30, 2026, Substack by Category sort rank_change stamped 2026-08-30T08:01:36 UTC and filled 14 climber rows on Technology window 3d, 18 rows on Sports 3d, 14 on News, 8 on Finance, and 7 on Faith & Spirituality, so a default sort rank pull at limit 25 never lists these mover fields.

Technology 3d climbers (ASCII rows, same as_of, baseline 2026-08-27T02:01:32 UTC):

Current rankTitleprev_rankrank_change
8LearnAIWithMe8072
24Nimble AF* Bytes7854
22Newcomer5836
27Sources6336
19Latent.Space5233
33Women + AI5421
14Angus's Substack3420
37AI for the 99%5114
12The Global Move2513
3Computer, Enhance!1512
4Pirate Wires117
1SemiAnalysis32
89Level Up with AI912

limit 50 on the same Technology 3d call still returned count 14. Raising limit does not invent extra climbers after the list is exhausted. lowercase technology resolved to the same Technology board and the same first five climbers.

Default sort rank on Technology filled SemiAnalysis at rank 1, Thinking Freely with Nita Farahany at rank 2, Computer, Enhance! at rank 3, Pirate Wires at rank 4, and Marcus on AI at rank 5. Rank 2 is absent from the 14-row climber list. Rank 1 is present with rank_change 2. The largest 3d climb sat at current rank 8. Level Up with AI sat at rank 89, prev_rank 91, rank_change 2. Default limit 25 never prints rank 89. The current-rank board for a category is documented on the Substack by Category API.

Sports 3d top ASCII climbers on the same stamp: Thinking About Thinking: Mostly Fantasy Football Newsletter at rank 16, prev_rank 66, rank_change 50; Data Dribble at 13 / 61 / +48; Small Craft Advisor at 54 / 99 / +45. Offset 18 then returned count 0 with the note that no items on the current board climbed versus the 3d baseline; everything held or fell. That note is the exhausted-page shape, not a claim that Sports had zero climbers.

Window 1d returns complete turnover on category boards

sort rank_change without a window is a 30d request, and Technology on this stamp set baseline_ts to null and returned count 0 with a note that no baseline snapshot sat within 10 days of 2026-07-31, so a copy-paste that only adds sort rank_change still looks empty.

WindowTypeCategorybaseline_tsCountNote
3dSubstack by CategoryTechnology2026-08-27T02:01:3214LearnAIWithMe +72 at rank 8
3dSubstack by CategorySports2026-08-27T02:01:3218Data Dribble +48 at rank 13
3dSubstack by CategoryNews2026-08-27T02:01:3214Grand Army Gazette +27 at rank 17
3dSubstack by CategoryFinance2026-08-27T02:01:328Across The Spread +28 at rank 22
3dSubstack by CategoryFaith & Spirituality2026-08-27T02:01:327Ad Aeternum +16 at rank 30
3dSubstack by CategoryClimate & Environment2026-08-27T00:01:300complete turnover; as_of 00:01:36
1dSubstack by CategoryTechnology2026-08-29T08:01:290complete turnover
1dSubstack by CategorySports2026-08-29T08:01:290complete turnover
7dSubstack by CategoryTechnology2026-08-25T20:01:410complete turnover
14dSubstack by CategoryTechnologynull0no baseline within 5 days of 2026-08-16
(omitted, defaults 30d)Substack by CategoryTechnologynull0no baseline within 10 days of 2026-07-31
1dSubstack(none)2026-08-29T02:01:3621Kirsten Bradley +52 at rank 45
3dSubstack(none)2026-08-27T02:01:3218Nancy & Wayne +42 at rank 8
7dSubstack(none)2026-08-26T02:01:3619Multipolarity +47 at rank 23
14dSubstack(none)null0no baseline within 5 days of 2026-08-16

Windows 1d and 7d on Technology still printed a baseline_ts. The empty body is a complete-turnover note, not missing history. Climate & Environment 3d used an older as_of of 2026-08-30T00:01:36 UTC and still returned complete turnover. Pass window 3d on in-sync category boards until a 1d overlap exists.

Sibling Substack Trending API on the same as_of filled 21 overall climbers on window 1d. Kirsten Bradley sat at rank 45, prev_rank 97, rank_change 52. Window 7d filled 19, led by Multipolarity at rank 23, prev_rank 70, rank_change 47. Those 1d and 7d fills do not transfer to type Substack by Category.

Women + AI appeared on both boards with different numbers. Technology 3d put it at rank 33, prev_rank 54, rank_change 21. type Substack 3d put it at rank 55, prev_rank 83, rank_change 28. Baltimore Film Breakdown sat at rank 36 on both Sports 3d and overall 3d, then split to prev_rank 53 / +17 on Sports versus prev_rank 67 / +31 on overall. Copying a category payload onto type Substack drops those gaps.

get_top_trends with type Substack by Category, category Technology, sort rank_change, and window 3d returns climber objects, and REST mode get_top_trends bills that call as one request per page, so one Technology 3d probe plus Sports 3d plus News 3d plus three empty longer Technology windows is six requests before any get_growth follow-up.

{
  "mode": "get_top_trends",
  "type": "Substack by Category",
  "category": "Technology",
  "sort": "rank_change",
  "window": "3d",
  "limit": 50
}
{
  "mode": "get_top_trends",
  "type": "Substack by Category",
  "category": "Sports",
  "sort": "rank_change",
  "window": "3d",
  "limit": 50
}
{
  "mode": "get_growth",
  "source": "google search",
  "keyword": "LearnAIWithMe",
  "percent_growth": ["7D", "30D", "3M", "12M"]
}

An unknown category returns HTTP 400 with 15 valid names: Art & Illustration, Business, Climate & Environment, Culture, Faith & Spirituality, Fashion & Beauty, Finance, Food & Drink, Health Politics, Music, News, Sports, Technology, U.S. Politics, World Politics. Comics returned error invalid_request. Names are the HTTP 400 list, not the longer table on llms.txt. Offset and limit still apply; details live on get_top_trends offset and limit.

Unfiltered rank_change reuses ranks across category boards

Omit category on sort rank_change and each row stays a category-local climb, so rank 10 can appear twice in one page while the category_path changes, which is the mixed-captain shape of the unfiltered current-rank board applied to movers.

Unfiltered 3d, limit 25, same as_of and baseline (first ASCII rows):

Current rankTitleprev_rankrank_changecategory_path
15Rupak's Substack9277Business
8LearnAIWithMe8072Technology
16Diary of a Speaker8872Business
4Circus Living7066Art & Illustration
30The Dad Briefs9262Food & Drink
28SwissCareer8759Business
14Ottolenghi7258Food & Drink
24Nimble AF* Bytes7854Technology

Rupak's Substack at Business +77 is a larger climb than any Technology row. It is absent from the Technology 14. Offset 25 on the same unfiltered call still returned count 25. Offset 50 returned 25. Offset 100 still returned 25, including Sound Bites at Music rank 2, prev_rank 8, rank_change 6. Default limit 25 hides those later pages. Do not treat this unfiltered page as a global newsletter mover list. Pair it with the filtered Substack Trending API board when a single 1-N ranking is required.

The sibling TikTok contract for category rank_change is documented on TikTok Hashtags by Category rank_change. That feed filled window 1d and emptied 3d. Substack by Category on this stamp is the inverse: 3d fills, 1d empties.

google search stays flat on the largest Technology climb

LearnAIWithMe can sit at Technology rank 8 with rank_change 72 while google search prints 0.0, so a Substack climb is not a substitute for a search-volume series.

Growth follow-ups on the same pull day (google search 261 weekly points, recent_date 2026-08-29):

KeywordLive rank contextWindowRecent vs baseline
LearnAIWithMeTechnology 3d rank 8 / +727D0.0 vs 0.0 (low quality, over 90 percent zeros)
LearnAIWithMeTechnology 3d rank 8 / +7212M0.0 vs 0.0
SemiAnalysisTechnology current rank 1 / +27D15.0 vs 17.0 (-11.76%)
SemiAnalysisTechnology current rank 1 / +230D15.0 vs 29.0 (-48.28%)
SemiAnalysisTechnology current rank 1 / +23M15.0 vs 30.0 (-50.0%)
Pirate WiresTechnology 3d rank 4 / +77D12.0 vs 19.0 (-36.84%)
Pirate WiresTechnology 3d rank 4 / +730D12.0 vs 0.0 (quote the two values, not 999999.0)
Data DribbleSports 3d rank 13 / +487D13.0 vs 14.0 (-7.14%)
Data DribbleSports 3d rank 13 / +4830D13.0 vs 47.0 (-72.34%)
Data DribbleSports 3d rank 13 / +483M13.0 vs 95.0 (-86.32%)
Women + AITechnology 3d rank 33 / +217D67.0 vs 71.0 (-5.63%)
Women + AITechnology 3d rank 33 / +213M67.0 vs 93.0 (-27.96%)
Rupak's Substackunfiltered 3d Business +777D0.0 vs 0.0 (low quality)
Nancy & Wayneoverall 3d rank 8 / +427D0.0 vs 7.0 (-100.0%)

SemiAnalysis can sit at Technology rank 1 while google search falls. Data Dribble can sit at Sports rank 13 while the 3M index drops from 95.0 to 13.0. Quote both payloads.

Common questions

Rows become objects with rank, keyword, prev_rank, rank_change, and category_path instead of a three-item rank-title-category array. On August 30, 2026 the Technology 3d window stamped 2026-08-30T08:01:36 UTC against baseline 2026-08-27T02:01:32 UTC and returned count 14. LearnAIWithMe sat at rank 8, prev_rank 80, rank_change 72. limit 50 on the same call still returned count 14.
sort rank_change defaults to window 30d. Technology with no window on this stamp set baseline_ts to null and returned count 0 with a note that no baseline sat within 10 days of 2026-07-31. Window 14d returned the same empty shape. Windows 1d and 7d did set a baseline_ts, then returned count 0 because the board turned over completely. Pass window 3d to get the only filled lookback on this type right now.
No. type Substack on the same as_of filled 21 climbers on window 1d and 19 on window 7d. Technology and Sports 1d both returned count 0 with a complete-turnover note. Window 14d was empty on both types with insufficient snapshot history. Poll each type. Do not copy a Substack 1d fill onto Substack by Category.
No. LearnAIWithMe sat at Technology 3d rank 8, prev_rank 80, rank_change 72, while google search sat at 0.0 across 7D, 30D, 3M, and 12M with a low-quality warning. SemiAnalysis sat at current rank 1 with rank_change 2 while google search sat at 15.0 versus 17.0 over 7 days. Quote the live rank and the growth payload separately.
get_top_trends is counted per feed type, per category, and per page, so one Technology 3d call is one request, separate from Sports 3d and from default sort rank. Offset 18 on the empty Sports 3d tail still billed a page and repeated the held-or-fell note. Free covers 100 requests per month and Starter is $19 for 1,000, with extra percent_growth windows inside one get_growth call still counting as a single request.