Today's story: Model Context Protocol
MCP becomes the universal adapter for AI's tool era
Model Context Protocol jumps to No. 7 as AI API SDKs and agents push a common interface for every app.
The numbers: Model Context Protocol is today's biggest mover: up 3 places to No. 7, installs at 1.23× the four-week pace. Model Context Protocol: 1.14× momentum across 3 signals. 6 of 21 trends are climbing and 3 are cooling. SpaceX Starship still leads the board at 2.4× (day 5).
Backfilled on October 7, 2026 using only data published by October 5, 2026.
Model Context Protocol, signal by signal
Each pedestal is one signal: its height is last week's typical day against the four weeks before, and the ring marks that normal. A lit pedestal is rising 1.15× or more.
Model Context Protocol is today's biggest mover: up 3 places to No. 7, installs at 1.23× the four-week pace.
Model Context Protocol: 12M npm installs a day on a typical day last week, against 9.7M the four weeks before
Bars are single days, the line is the 7-day average, the shaded band is the last 7 days and the dashed line is the four-week typical day (median).
Data through npm 10/03, Wikipedia 10/04, HN 10/04. Source: npm downloads API.
Model Context Protocol climbed three places to No. 7 with installs at 1.23× the four-week pace, hitting 12M/day. The protocol standardizes how AI agents reach external tools and data, turning every API into a plug-and-play resource. Its rise coincides with AI API SDKs at 21M/day installs, a 1.21× pace, and AI agents showing installs 1.25× attention. Developers are no longer wiring each app separately; they are adopting one interface that works across models and services.
The pattern mirrors Passkeys, which climbed to 1.0M/day installs at 1.31× momentum, replacing per-site passwords with a single authentication standard. Both MCP and Passkeys solve the same problem: reducing integration friction. As AI agents multiply, the cost of connecting them to every database, calendar, or payment system becomes the bottleneck. A common protocol turns that from a custom engineering task into a configuration step, which is why adoption accelerates even as attention stays flat.
Meanwhile, the cooling of smart glasses (0.67×) and robotaxis (0.73×) shows hardware hype fading while software standards rise. The board's six climbing trends are mostly protocols, SDKs, and batteries—enablers rather than end products. MCP's jump is the clearest signal: the next wave of adoption is not a single killer app but the layer that lets any AI talk to any tool. That is why it moved 3 places in one day.
Watch whether MCP installs keep outpacing AI agent attention, signaling that the protocol becomes the default gatekeeper for agent actions.
Watch next: Watch whether MCP installs keep outpacing AI agent attention, signaling that the protocol becomes the default gatekeeper for agent actions.
Column drafted by DeepSeek-V4-Flash from the numbers below; every figure checked against them.
The numbers
On a typical day last week, npm installs of @modelcontextprotocol/sdk reached 12M, 1.23× the 9.7M of a typical day in the four weeks before. Hacker News carried 7.4 stories a day with MCP in the title last week, against 6.6 a day in the four weeks before (1.12×). The Wikipedia article "Model Context Protocol" drew 1,538 readers on a typical day last week, 1.08× its four-week norm of 1,428.
Still No. 1 for day 5: SpaceX Starship at 2.4×, with Wikipedia readers at 2.4× (3,302 on a typical day against 1,378).
Quiet adoption: AI agents installs run at 1.25× their four-week pace while attention sits at 1.06×. Developers are adding it faster than people are reading about it.
Cooling fastest: Smart glasses at 0.67×, with Wikipedia readers down to 153 on a typical day from 230.
The long view: OpenAI Codex installs over the last four weeks are 117× the same four weeks a year earlier (3.4M a day against 29k).
Trend pages: SpaceX StarshipPasskeysStarlinkSmart glasses
Broad riseOne-day spike6 of 21 trends are climbing; SpaceX Starship leads at 2.4×
Momentum is a typical day (median) of the last 7 divided by a typical day of the 28 before, averaged across every signal with enough data. 1× means no change. Rings mark each signal.
Data through npm 10/03, Wikipedia 10/04, HN 10/04. Source: npm downloads API, Wikimedia pageviews, Hacker News search.
5 of 7 developer tools are being installed faster than they are being read about
Across: attention momentum (Wikipedia readers and Hacker News stories). Up: install momentum (npm). Both on log scales; the 1× lines split the corners.
Data through npm 10/03, Wikipedia 10/04, HN 10/04. Source: npm downloads API, Wikimedia pageviews, Hacker News search.
SpaceX Starship draws the most Wikipedia readers against its own normal: 2.4× over the last week
Each cell is one day's readers divided by that trend's four-week typical day. Brighter runs hot, darker runs cold.
Data through npm 10/03, Wikipedia 10/04, HN 10/04. Source: Wikimedia pageviews API, user agents only.
Every number on the board
| SpaceX Starship | Space | 2.4× | Breakout | – | – | 3,302 | 2.4× | 1.7 | 9.6×* | 0 |
|---|---|---|---|---|---|---|---|---|---|---|
| Passkeys | Web platform | 1.31× | Climbing | 1.0M | 1.31× | 86 | 0.97×* | 0.0 | 0×* | 0 |
| Starlink | Space | 1.24× | Climbing | – | – | 2,470 | 1.24× | 1.0 | 2.2×* | 1 |
| Solid-state batteries | Energy | 1.23× | Climbing | – | – | 489 | 1.23× | 0.0 | 0×* | 0 |
| AI API SDKs | AI agents | 1.21× | Climbing | 21M | 1.21× | – | – | – | – | 0 |
| OpenAI Codex | AI coding | 1.2× | Climbing | 3.9M | 1.16× | 53 | 1.38×* | 5.3 | 1.25× | 0 |
| Model Context Protocol | AI agents | 1.14× | Steady | 12M | 1.23× | 1,538 | 1.08× | 7.4 | 1.12× | 0 |
| WebGPU | Web platform | 1.14× | Steady | 1.7M | 1.23× | 135 | 1.06× | 1.0 | 1.47×* | 0 |
| AI agents | AI agents | 1.12× | Steady | 755k | 1.25× | 962 | 0.82× | 15 | 1.38× | 1 |
| Claude Code | AI coding | 1.1× | Steady | 2.4M | 1.08× | – | – | 7.4 | 1.12× | 0 |
| Bun runtime | Web platform | 1.1× | Steady | 758k | 1.19× | 244 | 1.01× | 0.9 | 0.92×* | 0 |
| Small modular reactors | Energy | 1.1× | Steady | – | – | 510 | 1.1× | 0.0 | 0×* | 0 |
| Gemini CLI | AI coding | 1.09× | Steady | 70k | 1.09× | – | – | 0.0 | n/a* | 0 |
| Local LLMs | AI agents | 1.08× | Steady | 192k | 1.13× | 721 | 1.03× | 1.1 | 1.88×* | 0 |
| Heat pumps | Energy | 1.08× | Steady | – | – | 510 | 1.08× | 0.0 | 0×* | 0 |
| Vibe coding | AI coding | 1.02× | Steady | – | – | 1,734 | 1.2× | 1.0 | 0.88× | 0 |
| Humanoid robots | Robots | 0.91× | Steady | – | – | 338 | 0.91× | 0.9 | 1.09×* | 0 |
| eVTOL air taxis | Space | 0.89× | Steady | – | – | 182 | 0.89× | 0.0 | 0×* | 0 |
| Quantum computing | Devices | 0.81× | Cooling | – | – | 1,721 | 0.98× | 2.1 | 0.67× | 0 |
| Robotaxis | Robots | 0.73× | Cooling | – | – | 184 | 0.73× | 0.4 | 1×* | 2 |
| Smart glasses | Devices | 0.67× | Cooling | – | – | 153 | 0.67× | 0.3 | 0.4×* | 0 |
| Vehicle-to-grid | Energy | n/a | Not enough data | – | – | 64 | 1.03×* | 0.0 | 0×* | 1 |
Method and sources
Every trend gets up to three free, daily signals. Installs: npm registry downloads for the packages named on each trend page (developer adoption; it counts CI and mirrors too, so it measures direction, not people). Attention: English Wikipedia readers per day, human agents only, and Hacker News stories with the term in the title. For each signal we divide a typical day (the median) of the last 7 complete days by a typical day of the 28 days before. Medians keep one launch-day spike from carrying a whole week; both windows are whole weeks, so weekend dips sit equally in both. Momentum is the geometric mean of the signals that clear a minimum base. npm reports some days as zero for every package at once; we treat those as missing.
The column is drafted by a free language model from the day's computed facts only, and is published only if every number in it matches those facts; otherwise the computed paragraphs stand alone. Theme labels come from Jev (or its Cloudflare Clef fallback, as noted) choosing from a fixed list of ten.
Stages: Breakout 1.5× and up, Climbing 1.15×, Steady 0.87× to 1.15×, Cooling below. Raw data: /api/trends/2026-10-05.json.