Substack rank_change via get_top_trends

Pass sort rank_change on type Substack to list newsletter climbers, then read empty 30d notes against filled 1d, 3d, and 7d windows from one account.

Live data as of 2026-08-30

get_top_trends with type Substack, sort rank_change, and an explicit window returns newsletter titles that moved up versus a stored snapshot. Default window is 30d. On the August 30, 2026 scan stamped 2026-08-30T12:01:43 UTC that default set baseline_ts to null and returned count 0, with a note that no snapshot sat within 10 days of 2026-07-31. Windows 1d, 3d, and 7d filled. 1d closed at 26 climbers once limit was raised to 50. 3d returned 19. 7d returned 20. Rows are objects with rank, keyword, prev_rank, and rank_change, not the three-item arrays that sort rank returns. Overall objects on this stamp had no category_path.

Default window 30d is empty on this feed

sort rank_change without a window is a 30d lookback. The August 30, 2026 call with type Substack and no window (and the explicit window 30d call) both returned count 0, data as an empty list, and baseline_ts null. The note named a target of 2026-07-31 and a 10-day tolerance. Window 14d was empty the same way, targeting 2026-08-16 within 5 days.

The filled windows used real baselines. 1d used 2026-08-29T02:01:36 UTC. 3d used 2026-08-27T02:01:32 UTC. 7d used 2026-08-26T02:01:36 UTC, about four days before the 12:01 stamp, not seven. The API accepted a snapshot inside its window tolerance. Quote baseline_ts next to window. Do not assume a 7d string means a calendar week.

REST shape for the window that actually filled:

{"mode": "get_top_trends", "type": "Substack", "sort": "rank_change", "window": "1d", "limit": 50}

MCP is the same fields on get_top_trends. type is required on MCP. REST can omit type to mix feeds; keep type Substack when the point is this board.

limit 25 truncates the 1d climber list

Default limit is 25 and max is 200. On this stamp the 1d window returned count 25 at limit 25 and count 26 at limit 50. The 26th row was The Style Edit at rank 5, prev_rank 6, rank_change 1. Offset 25 with limit 25 returned that single row. Offset 26 returned count 0, baseline_ts still 2026-08-29T02:01:36 UTC, and a note that no items on the current board climbed versus the 1d baseline; everything held or fell.

That held-or-fell note is a pagination stop, not a missing snapshot. 3d with limit 50 still returned count 19. Offset 19 on 3d returned the same held-or-fell note. 7d returned count 20 at offset 0 and count 0 at offset 20, again with baseline_ts set. Compare that to 14d and 30d, where baseline_ts is null and the note names insufficient snapshot history. Read both fields before retrying a longer window. Each get_top_trends page counts as one request. The get_top_trends offset and limit page covers feed-length caps on other types.

1d leaders on the filled 50-row call, skipping titles that need non-ASCII glyphs: Bella Oldham | Avocado Bella 24 / 96 / +72. ME THANKY 13 / 59 / +46. Babylon Burns 44 / 88 / +44. Kirsten Bradley 54 / 97 / +43. The Neter sat at rank 2, prev_rank 4, rank_change 2, matching current-board rank 2 on the same as_of.

Category 3d complete turnover on the same as_of

Five Substack by Category 3d calls on 2026-08-30T12:01:43 UTC returned count 0 with a complete-turnover note: Sports, Technology, News, Finance, and Faith & Spirituality. Each set baseline_ts to 2026-08-27T16:01:35 UTC. Overall Substack 3d on that same as_of used 2026-08-27T02:01:32 UTC and returned 19 climbers. The category boards and the overall board do not share a snapshot clock, and a complete-turnover note is not the same as a missing-baseline note.

Overall 3d, ASCII titles only: FOOD YOU WANT TO EAT 21 / 68 / +47. Backstage Pass 56 / 93 / +37. Tretas Club 23 / 59 / +36. Nancy & Wayne 15 / 50 / +35. Baltimore Film Breakdown 46 / 67 / +21. The Style Edit 5 / 10 / +5. Unhappy Place 18 / 20 / +2. limit 50 did not invent extra 3d rows past 19.

Current sort rank on type Substack, limit 10, same stamp: rank 1 Inside Wolves with Liam Keen and Nathan Judah. Rank 2 The Neter. Rank 3 InsideNUFC. Rank 4 Your Brain On Mom. Rank 5 The Style Edit. Rank 6 The Moynihan Report. Rank 7 Callie Coles. Rank 8 Spencer Pratt. Rank 9 Aimee Song. Rank 10 Laurah Wheaton. The Neter is a 1d climber. Your Brain On Mom is a 7d climber (4 / 15 / +11). The Style Edit appears on both 1d and 3d. Spencer Pratt at rank 8 is not on the 1d or 3d climber lists. The overall Substack rank board is the sort rank view. The Substack by Category rank_change page is the split-board view. Pass an official category name there. Type Substack takes no category.

Rank climb is not google search growth

get_growth with source google search and recent_date 2026-08-29 (261 weekly points) did not track the Substack 3d leader. FOOD YOU WANT TO EAT sat at 77.0 versus 80.0 over 7 days (minus 3.75 percent), 77.0 versus 67.0 over 30 days (plus 14.93 percent), and 77.0 versus 81.0 over 3 months (minus 4.94 percent). The newsletter climbed 47 places on Substack 3d while the 7-day search index fell.

Bella Oldham, the 1d leader at +72, returned 13.0 versus 14.0 over 7 days (minus 7.14 percent) and 13.0 versus 100.0 over 3 months (minus 87.0 percent). The 30-day row was 13.0 versus 0.0. data_quality was low, with a warning that the series is over 90 percent zeros. Quote the two index values. Do not publish the 999999.0 percent the API emits when baseline_value is 0.0.

Other follow-ups on the same recent_date. The Neter, current rank 2 and a 1d climber of 2 places, sat at 53.0 versus 59.0 over 7 days (minus 10.17 percent). Your Brain On Mom, 7d rank 4 from prev_rank 15, sat at 50.0 versus 41.0 over 7 days (plus 21.95 percent) and 50.0 versus 77.0 over 3 months (minus 35.06 percent). The Style Edit sat at 30.0 versus 25.0 over 7 days (plus 20.0 percent) and 30.0 versus 67.0 over 3 months (minus 55.22 percent). Callie Coles, 7d rank 7 from prev_rank 8, sat at 52.0 versus 52.0 over both 7 days and 30 days. Multipolarity, 7d rank 39 from prev_rank 70, sat at 11.0 versus 11.0 over 7 days and 11.0 versus 25.0 over 30 days (minus 56.0 percent). Nancy & Wayne, 3d rank 15 from prev_rank 50, sat at 7.0 versus 8.0 over 7 days with data_quality low. Kirsten Bradley, 1d rank 54 from prev_rank 97, sat at 19.0 versus 12.0 over 7 days (plus 58.33 percent), also low quality. Baltimore Film Breakdown sat at 0.0 versus 7.0 over 7 days (minus 100.0 percent), low quality. Inside Wolves (the rank-1 search string, not the full title) sat at 56.0 versus 54.0 over 7 days (plus 3.7 percent) and 56.0 versus 95.0 over 30 days (minus 41.05 percent); that keyword collides with a football club nickname. Spencer Pratt at overall rank 8 sat at 2.0 versus 1.0 over 7 days and 2.0 versus 33.0 over 3 months (minus 93.94 percent).

get_growth is one request per source plus keyword, with every percent_growth period included. Free covers 100 requests per month. Starter is $19 for 1,000. Pro is $49 for 5,000. Business is $199 for 25,000. Check current pricing before locking a production budget.

sort rank still returns [rank, title] arrays. sort rank_change returns objects. Mixing the two parsers is the usual client bug. Empty count 0 is three different states: missing baseline (14d, 30d), complete turnover (category 3d on this stamp), or held-or-fell after the last climber (1d offset 26, 3d offset 19, 7d offset 20). The note string tells them apart. Window 7d filling against an August 26 baseline will keep mattering after this hour's titles rotate off the board.

Common questions

Rows become objects with rank, keyword, prev_rank, and rank_change. Overall Substack objects on this scan had no category_path field. On August 30, 2026 the 1d window stamped 2026-08-30T12:01:43 UTC against baseline 2026-08-29T02:01:36 UTC and closed at count 26 when limit was 50. Bella Oldham | Avocado Bella sat at rank 24, prev_rank 96, rank_change 72.
sort rank_change defaults to window 30d. On this stamp the 30d call 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, targeting 2026-08-16 within 5 days. Pass window 1d, 3d, or 7d to get a filled lookback on this feed right now.
No. The 1d window returned count 25 at limit 25 and count 26 at limit 50. Offset 25 then returned the missing row, The Style Edit at rank 5, prev_rank 6, rank_change 1. Offset 26 returned count 0 with a held-or-fell note and baseline_ts still set. Raise limit or page once before treating an empty page as a missing snapshot.
Not on this stamp. Sports, Technology, News, Finance, and Faith & Spirituality 3d calls all returned count 0 with a complete-turnover note and baseline_ts 2026-08-27T16:01:35 UTC. Overall Substack 3d on the same as_of used baseline 2026-08-27T02:01:32 UTC and returned count 19. Read the note field. Do not copy a category count onto type Substack.
No. FOOD YOU WANT TO EAT led the overall 3d list at rank 21, prev_rank 68, rank_change 47, while google search sat at 77.0 versus 80.0 over 7 days (minus 3.75 percent), recent_date 2026-08-29. Bella Oldham climbed 72 places on 1d with a 13.0 versus 14.0 7-day index and a low-quality series. Pair get_top_trends with get_growth instead of inferring demand from rank_change.