Compare exact npm package names and exact PyPI project names in one Trends MCP account, using weekly download volumes and the dates those weeks actually close.
Live data as of 2026-08-24
An npm vs PyPI downloads API on Trends MCP is two sources on one key: npm for exact npm package names, python for exact PyPI project names. On the August 24, 2026 pull, react closed the week of 2026-08-22 at index 100.0 with 170,203,436 downloads across 261 weekly points, while the open week of 2026-08-29 sat at 0.0. pandas on python closed the week of 2026-08-10 at index 93.1 with 187,765,094 downloads across 260 points, up 12.58% from 82.7 (volume 168,389,688) on 2026-07-13. requests on python sat at index 100.0 with 443,005,393 downloads that same week. The free plan includes 100 requests per month.
From the Trends MCP pull of August 24, 2026, npm weekly series for react still had an open week at 2026-08-29 (volume 0), so the last complete npm week in this file is 2026-08-22, while python series for pandas, requests, numpy, openai, pydantic, and fastapi all used recent_date 2026-08-10.
| Registry | Package | Week | Index | Weekly downloads |
|---|---|---|---|---|
| npm | react | 2026-08-22 | 100.0 | 170,203,436 |
| npm | react | 2026-08-08 | 95.6 | 162,776,333 |
| npm | react | 2026-08-01 | 95.4 | 162,348,333 |
| npm | lodash | 2026-08-01 | 96.6 | 165,786,732 |
| npm | typescript | 2026-08-01 | 95.7 | 257,651,265 |
| npm | next | 2026-08-01 | 100.0 | 54,973,809 |
| npm | openai | 2026-08-01 | 93.9 | 32,375,676 |
| python | pandas | 2026-08-10 | 93.1 | 187,765,094 |
| python | requests | 2026-08-10 | 100.0 | 443,005,393 |
| python | numpy | 2026-08-10 | 95.0 | 286,323,325 |
| python | pydantic | 2026-08-10 | 99.0 | 286,050,804 |
| python | fastapi | 2026-08-10 | 99.0 | 148,979,438 |
| python | openai | 2026-08-10 | 95.5 | 109,503,647 |
typescript at 257,651,265 weekly downloads on 2026-08-01 outran react's 162,348,333 on the same npm week. requests at 443,005,393 on 2026-08-10 outran pandas at 187,765,094 on the python week. Those are not the same seven days. A cross-registry ratio that ignores the week-ending date is a made-up statistic.
Preset npm windows in this pull used recent_date 2026-08-29, and that week was still at volume 0.0, so get_growth reported -100% for react, lodash, typescript, next, and openai even though the prior complete weeks were large.
react get_time_series (261 points from 2021-09-04 through 2026-08-29) shows the drop is a trailing zero, not a collapse. 2026-08-22 is 100.0 / 170,203,436. 2026-08-15 is 67.9 / 115,573,856. 2026-08-08 is 95.6 / 162,776,333. 2026-07-25 is 95.0 / 161,625,306. 2026-05-30 is 75.8 / 128,997,198. 2025-08-30 is 29.1 / 49,473,446. Against 2025-08-23 at 25.6 / 43,546,734, the 2026-08-22 week is several times larger. Publish the dated volume. Do not publish the -100% print.
lodash get_growth baseline_volume on 2026-08-01 was 165,786,732 (index 96.6), on 2026-05-30 was 148,701,673 (index 86.6), and on 2025-08-30 was 74,195,295 (index 43.2). next sat at 54,973,809 (index 100.0) on 2026-08-01, 39,436,457 (index 71.7) on 2026-05-30, and 13,230,887 (index 24.1) on 2025-08-30. openai on npm sat at 32,375,676 (index 93.9) on 2026-08-01 versus 6,318,725 (index 18.3) on 2025-08-30. For package-level npm history without the PyPI side, see the npm Trends API.
One get_growth call per registry plus keyword is enough for several windows, because percent_growth periods share the request, while mixing npm and python in a single source string is not how these two registries are queried.
{
"keyword": "react",
"source": "npm",
"percent_growth": ["30D", "3M", "12M"]
}
{
"keyword": "pandas",
"source": "python",
"percent_growth": ["30D", "3M", "6M", "12M"]
}
npm keywords are case-sensitive package names (react, lodash, typescript, next, openai, @babel/core). python keywords are PyPI project names (pandas, requests, numpy, openai, pydantic, fastapi). Display names and import paths miss. get_time_series for react returned 261 weekly points. python growth metadata for pandas, requests, numpy, openai, pydantic, and fastapi reported 260 points. When npm get_growth recent_volume is 0.0, read get_time_series and take the last week with volume above zero. For PyPI-only volume and the missing top-packages board, see the PyPI package downloads API.
python recent_date in this pull was 2026-08-10 for every successful project, about two weeks behind the August 24 request, while npm's last complete week was 2026-08-22 and its open week was dated 2026-08-29, five days ahead of the pull day.
pandas rose 12.58% over 30 days (93.1 from 82.7, volumes 187,765,094 from 168,389,688), 19.36% over 3 months from 78.0 (volume 159,668,659 on 2026-05-11), 59.42% over 6 months from 58.4 (volume 122,957,370 on 2026-02-09), and 128.75% over 12 months from 40.7 (volume 89,972,939 on 2025-08-11). requests rose 14.68% over 30 days from 87.2 (volume 389,484,929), 15.07% over 3 months from 86.9 (volume 388,093,628), 70.94% over 6 months from 58.5 (volume 269,649,289), and 163.16% over 12 months from 38.0 (volume 184,043,469). numpy sat at 95.0 / 286,323,325, up 13.37% over 30 days, 32.5% over 3 months, and 175.36% over 12 months from 34.5 / 114,381,653. pydantic sat at 99.0 / 286,050,804, up 15.93% over 30 days, 7.96% over 3 months, and 222.48% over 12 months from 30.7 / 90,159,377. fastapi sat at 99.0 / 148,979,438, up 17.02% over 30 days from 84.6 / 127,435,423, 17.16% over 3 months, and 475.58% over 12 months from 17.2 / 26,515,604. openai on python sat at 95.5 / 109,503,647, up 26.49% over 30 days from 75.5 / 86,559,018, 53.78% over 3 months from 62.1 / 71,239,852, and 372.77% over 12 months from 20.2 / 23,146,365.
openai the npm package at 32,375,676 on 2026-08-01 and openai the PyPI project at 109,503,647 on 2026-08-10 are not a single product. The names match. The registries, weeks, and install bases do not. GitHub Trending Repos on 2026-08-24T02:01:35 UTC listed 18 rows with openai/codex at rank 1 and anthropics/claude-plugins-community at rank 15. That board is repo heat, not weekly installs. For repo ranks beside these download series, see the GitHub Trending Repos API.
Weekly download totals are not unique users, production installs, or GitHub stars, which is why typescript at 257,651,265 on 2026-08-01 can exceed react at 162,348,333 on the same npm week without proving TypeScript is "more popular" than React in product terms.
CI systems retry. Monorepos install the same package many times. Mirror traffic differs by registry. python volume in this pull lags about two weeks (recent_date 2026-08-10). npm get_growth can point at a future week_end (2026-08-29) with zeros. The 2026-08-15 react week at 67.9 / 115,573,856 sits well below 2026-08-08 and 2026-08-22; treat mid-month dips as possibly incomplete rather than as a demand crash. There is no get_top_trends type for either registry, so a "top 10 packages" page cannot be built from a live board. Keyword format is strict: @scoped npm names need the exact scope, and PyPI names follow the project slug on pypi.org, not the import.
The free plan includes 100 requests per month with no credit card, Starter is $19 per month for 1,000 requests, Pro is $49 for 5,000, and Business is $199 for 25,000, and each get_growth or get_time_series call counts as one request even when several percent_growth windows ride along.
Annual billing saves 20% on the paid plans. Check current pricing before locking a production budget. For npm series without the PyPI comparison, see npm trends.
FAQ