Most crowded categories in the agent ecosystem
Saturation ranking of 454 taxonomy segments and 111 integrations: size, launch momentum and concentration of attention.
Executive summary
Agent Commerce is the largest top-level category with 3,621 products; AI Coding Agents follows with 3,366.
69 of 454 taxonomy segments launch faster than the catalogue as a whole (labelled outpacing or emerging); 176 lag it or are slowing.
In 139 of 441 segments with at least five repositories, a single project holds half or more of all GitHub stars.
The most crowded specific capability is SEO Research with 1,073 products, 735 of them launched in the last 90 days.
Among integrations, YouTube is mentioned by the most products (574).
Every figure above is computed from the database DER
Methodology
How this report was produced, and its limits
Generated on 2026-11-21 from the AgentSidekick database. The latest completed ingestion step finished at 2026-11-21 13:23 UTC. Tables queried: market_segments, taxonomy_nodes, integrations. The catalogue holds 73,820 active products, of which 9,469 meet the data-completeness bar and have a profile page.
Segments were last calculated at 2026-11-21 13:23 UTC.
Classification method: product type comes from the source (registry entries are MCP servers by definition) or is inferred from repository topics; categories, capabilities, integrations and industries come from keyword rules in the editable taxonomy, so they are derived, not checked by a person. Curated records were researched against the official sources listed on each record.
Known limits:
- Segment size counts every active product matched to the segment, profiled or not. A product can sit in several segments.
- Top share = the share of the segment's GitHub stars held by its most-starred project; HHI = the sum of squared star shares (0 to 1; above 0.25 is conventionally "highly concentrated"). Both only use products with a repository, and stars are attention, not market share.
- Labels compare a segment's size with peers at the same level (80th percentile and up = crowded, 30th and below = thin) and its launch momentum with the catalogue's: momentum = 50 + 50 × tanh(ln(segment ratio ÷ catalogue ratio)), where ratio = (launches in the last 90 days + 1) ÷ (launches in the 90 before + 1). 62 or more = outpacing the catalogue, 38 or less = lagging or slowing. Fewer than five launches across both windows = no momentum verdict. Because the whole catalogue is growing fast, a segment can add products every month and still be labelled lagging.
- Launch dates are the first publish date in the source registry (official MCP Registry) or the GitHub repository creation date. They are not product announcements, and a re-published entry can look like a new launch.
- GitHub stars measure attention on a repository, not installs, active users or revenue. Several products can share one repository.
- Only 1 daily metrics snapshot(s) exist in product_metrics, so growth is measured from launch counts, not from changes in stars or downloads.
- Categories, capabilities and integrations are assigned by keyword rules over names and descriptions (derived). Products that do not name a capability are missed, and generic words can produce false matches.
Note. Claim labels: source = stated by the publisher or registry; curated = checked by us against cited sources; derived = computed by a documented rule; heuristic = a rule of thumb with no measurement behind it. Full method: methodology.
How segments are labelled
Every taxonomy node and integration is scored on two axes: size relative to its peers, and launch momentum. "Crowded" means the segment is in the top fifth by size at its level; the suffix says whether its launches are growing faster or slower than launches across the whole catalogue. 17 segments are crowded and still outpacing the catalogue, and 120 are thin. A crowded label is a description of supply, not a statement that there is no room: a large segment with a dispersed star distribution can still reward a better product.
Taxonomy segments by labelDER
Unit: segments. Top 9 shown.
Show the data as a table
| segments | |
|---|---|
| Lagging the catalogue | 125 |
| Thin | 120 |
| Established | 62 |
| Crowded, slowing | 51 |
| Outpacing the catalogue | 35 |
| Crowded | 26 |
| Emerging | 17 |
| Crowded, outpacing the catalogue | 17 |
| Empty | 1 |
Top-level categories (level 2)
Agent Commerce is the largest level-2 segment with 3,621 products (2,284 vs 782 launches in the last two 90-day windows). 1 of the 20 largest are crowded and launching faster than the catalogue as a whole. The most concentrated among them is Finance & Accounting, where the leading project holds 33% of GitHub stars (HHI 0.16). The least concentrated is Agent Frameworks & Orchestration (top share 5%, HHI 0.02), where no single project dominates.
Largest level-2 categoriesDER
| Segment | Products | With repo | New 90d | Prev 90d | Top share | HHI | Label |
|---|---|---|---|---|---|---|---|
| Agent Commerce | 3,621 | 1,709 | 2,284 | 782 | 17% | 0.08 | Crowded |
| AI Coding Agents | 3,366 | 2,824 | 1,334 | 695 | 7% | 0.02 | Crowded, slowing |
| Communication & Messaging | 2,687 | 1,387 | 1,743 | 468 | 33% | 0.14 | Crowded |
| Public & Real-world Data | 2,011 | 1,371 | 1,113 | 622 | 23% | 0.11 | Crowded, slowing |
| Social & Streaming Platforms | 2,001 | 1,013 | 1,391 | 339 | 22% | 0.08 | Crowded, outpacing the catalogue |
| Marketing | 1,944 | 1,161 | 1,289 | 358 | 23% | 0.10 | Crowded |
| Cloud Providers | 1,777 | 1,277 | 903 | 342 | 13% | 0.04 | Crowded |
| Retrieval & RAG | 1,756 | 1,485 | 598 | 420 | 5% | 0.02 | Crowded, slowing |
| Crypto & Blockchain | 1,614 | 974 | 853 | 367 | 21% | 0.09 | Crowded, slowing |
| Files & Documents | 1,613 | 963 | 958 | 333 | 20% | 0.07 | Established |
| Finance & Accounting | 1,600 | 578 | 1,197 | 270 | 33% | 0.16 | Outpacing the catalogue |
| Agent Security | 1,481 | 1,110 | 650 | 368 | 9% | 0.03 | Lagging the catalogue |
| Web Data | 1,470 | 1,064 | 967 | 258 | 20% | 0.08 | Established |
| Agent Runtime & Deployment | 1,461 | 1,170 | 614 | 394 | 9% | 0.04 | Lagging the catalogue |
| DevOps & SRE | 1,392 | 1,147 | 518 | 313 | 4% | 0.02 | Lagging the catalogue |
| Research & Knowledge | 1,279 | 921 | 598 | 291 | 11% | 0.04 | Lagging the catalogue |
| E-commerce | 1,242 | 655 | 777 | 268 | 20% | 0.10 | Established |
| Generative Media | 1,232 | 931 | 705 | 286 | 10% | 0.04 | Lagging the catalogue |
| Markets & Trading | 1,209 | 737 | 684 | 297 | 28% | 0.14 | Lagging the catalogue |
| Agent Frameworks & Orchestration | 1,192 | 1,040 | 311 | 204 | 5% | 0.02 | Lagging the catalogue |
Subcategories (level 3)
Agent Payments & Monetization is the largest level-3 segment with 3,621 products (2,284 vs 782 launches in the last two 90-day windows). 6 of the 20 largest are crowded and launching faster than the catalogue as a whole. The most concentrated among them is Sales Automation, where the leading project holds 50% of GitHub stars (HHI 0.28). The least concentrated is Retrieval-Augmented Generation (top share 6%, HHI 0.02), where no single project dominates.
Largest level-3 subcategoriesDER
| Segment | Products | With repo | New 90d | Prev 90d | Top share | HHI | Label |
|---|---|---|---|---|---|---|---|
| Agent Payments & Monetization | 3,621 | 1,709 | 2,284 | 782 | 17% | 0.08 | Crowded |
| 1,718 | 834 | 1,169 | 300 | 30% | 0.14 | Crowded | |
| Marketing Automation | 1,670 | 971 | 1,135 | 284 | 29% | 0.12 | Crowded, outpacing the catalogue |
| Finance Automation | 1,600 | 578 | 1,197 | 270 | 33% | 0.16 | Crowded, outpacing the catalogue |
| Context Engineering | 1,466 | 1,292 | 480 | 350 | 9% | 0.03 | Crowded, slowing |
| Social Networks | 1,463 | 746 | 1,013 | 251 | 15% | 0.06 | Crowded, outpacing the catalogue |
| Document Processing | 1,405 | 815 | 887 | 279 | 29% | 0.12 | Crowded |
| Blockchain Data | 1,352 | 828 | 716 | 317 | 22% | 0.09 | Crowded, slowing |
| Edge & Serverless Platforms | 1,280 | 934 | 724 | 238 | 19% | 0.07 | Crowded |
| Retrieval-Augmented Generation | 1,193 | 1,027 | 394 | 277 | 6% | 0.02 | Crowded, slowing |
| Public Datasets | 1,126 | 849 | 668 | 347 | 18% | 0.10 | Crowded, slowing |
| Code Review | 1,094 | 946 | 546 | 225 | 35% | 0.24 | Crowded, slowing |
| Chat & Messaging | 1,041 | 580 | 624 | 181 | 38% | 0.18 | Crowded |
| Web Scraping | 922 | 744 | 633 | 147 | 24% | 0.11 | Crowded, outpacing the catalogue |
| Video & Audio Platforms | 879 | 466 | 616 | 157 | 32% | 0.13 | Crowded |
| Online Store Management | 859 | 463 | 594 | 148 | 16% | 0.06 | Crowded, outpacing the catalogue |
| Memory & State | 850 | 727 | 320 | 230 | 12% | 0.05 | Crowded, slowing |
| Operations & Supply Chain | 814 | 528 | 433 | 231 | 40% | 0.18 | Crowded, slowing |
| Sales Automation | 813 | 450 | 550 | 136 | 50% | 0.28 | Crowded, outpacing the catalogue |
| Market Data | 788 | 473 | 436 | 200 | 37% | 0.19 | Crowded, slowing |
Specific capabilities (level 4)
SEO Research is the largest capability segment with 1,073 products (735 vs 197 launches in the last two 90-day windows). 5 of the 20 largest are crowded and launching faster than the catalogue as a whole. The most concentrated among them is DeFi Protocols, where the leading project holds 87% of GitHub stars (HHI 0.77). The least concentrated is Multi-agent Coordination (top share 8%, HHI 0.03), where no single project dominates.
Largest capabilitiesDER
| Segment | Products | With repo | New 90d | Prev 90d | Top share | HHI | Label |
|---|---|---|---|---|---|---|---|
| SEO Research | 1,073 | 659 | 735 | 197 | 37% | 0.18 | Crowded |
| PDF Processing | 1,068 | 618 | 712 | 200 | 39% | 0.21 | Crowded |
| Invoice Processing | 927 | 273 | 722 | 141 | 72% | 0.56 | Crowded, outpacing the catalogue |
| Agent Memory | 804 | 699 | 312 | 223 | 13% | 0.05 | Crowded, slowing |
| YouTube Content | 697 | 374 | 495 | 122 | 12% | 0.07 | Crowded, outpacing the catalogue |
| Code Quality | 596 | 564 | 356 | 122 | 21% | 0.08 | Crowded |
| Domains & DNS | 573 | 302 | 395 | 94 | 38% | 0.24 | Crowded, outpacing the catalogue |
| Consumer Messaging | 561 | 318 | 355 | 102 | 57% | 0.35 | Crowded |
| Cloudflare Platform | 545 | 513 | 281 | 120 | 42% | 0.22 | Crowded, slowing |
| Remote MCP Hosting | 515 | 372 | 273 | 151 | 44% | 0.23 | Crowded, slowing |
| Multi-agent Coordination | 489 | 400 | 122 | 78 | 8% | 0.03 | Crowded, slowing |
| CRM Updates | 472 | 215 | 310 | 85 | 56% | 0.35 | Crowded |
| Semantic Search | 409 | 348 | 103 | 110 | 13% | 0.06 | Crowded, slowing |
| Headless Browsing | 393 | 340 | 181 | 87 | 12% | 0.06 | Crowded, slowing |
| Tax Compliance | 374 | 131 | 294 | 57 | 36% | 0.23 | Crowded, outpacing the catalogue |
| DeFi Protocols | 360 | 269 | 150 | 94 | 87% | 0.77 | Crowded, slowing |
| OCR | 358 | 232 | 228 | 77 | 29% | 0.13 | Crowded |
| Instruction Management | 348 | 310 | 73 | 58 | 36% | 0.18 | Crowded, slowing |
| Webhooks & Triggers | 330 | 180 | 193 | 82 | 24% | 0.18 | Crowded, slowing |
| Campaign Management | 324 | 164 | 228 | 43 | 35% | 0.27 | Crowded, outpacing the catalogue |
What this does not show: whether products in a capability are interchangeable. Two products tagged with the same capability may differ completely in quality, hosting model and maturity; the label only counts them.
Integrations
YouTube is the largest integration segment with 574 products (409 vs 98 launches in the last two 90-day windows). 7 of the 20 largest are crowded and launching faster than the catalogue as a whole. The most concentrated among them is Google Maps, where the leading project holds 72% of GitHub stars (HHI 0.56). The least concentrated is Azure (top share 15%, HHI 0.08), where no single project dominates.
Integrations mentioned by the most productsDER
| Segment | Products | With repo | New 90d | Prev 90d | Top share | HHI | Label |
|---|---|---|---|---|---|---|---|
| YouTube | 574 | 251 | 409 | 98 | 17% | 0.11 | Crowded, outpacing the catalogue |
| GitHub | 552 | 325 | 245 | 98 | 20% | 0.09 | Crowded, slowing |
| 420 | 196 | 287 | 78 | 24% | 0.13 | Crowded | |
| Solana | 357 | 158 | 206 | 86 | 61% | 0.45 | Crowded, slowing |
| PostgreSQL | 287 | 194 | 140 | 61 | 20% | 0.12 | Crowded, slowing |
| 285 | 132 | 200 | 48 | 32% | 0.22 | Crowded, outpacing the catalogue | |
| Shopify | 228 | 85 | 163 | 38 | 53% | 0.35 | Crowded, outpacing the catalogue |
| Android | 228 | 155 | 136 | 32 | 15% | 0.08 | Crowded, outpacing the catalogue |
| Telegram | 200 | 114 | 130 | 33 | 41% | 0.25 | Crowded |
| 196 | 106 | 112 | 36 | 38% | 0.22 | Crowded | |
| Stripe | 189 | 56 | 82 | 75 | 39% | 0.32 | Crowded, slowing |
| AWS | 162 | 100 | 43 | 35 | 17% | 0.08 | Crowded, slowing |
| WordPress | 155 | 82 | 112 | 19 | 70% | 0.52 | Crowded, outpacing the catalogue |
| Google Ads | 153 | 74 | 115 | 18 | 47% | 0.44 | Crowded, outpacing the catalogue |
| Gmail | 126 | 75 | 63 | 25 | 39% | 0.28 | Crowded, slowing |
| Azure | 124 | 67 | 38 | 15 | 15% | 0.08 | Crowded, slowing |
| Kubernetes | 120 | 92 | 38 | 26 | 21% | 0.11 | Crowded, slowing |
| Google Maps | 119 | 49 | 98 | 10 | 72% | 0.56 | Crowded, outpacing the catalogue |
| Slack | 114 | 72 | 44 | 16 | 38% | 0.21 | Crowded |
| MySQL | 112 | 79 | 45 | 27 | 28% | 0.17 | Crowded, slowing |
Integration counts come from product names and descriptions that mention the service. Official vendor-built servers and community wrappers count the same here; see the integration coverage report for demand against supply.
Recommendations
Editorial: the judgement of AgentSidekick, drawn from the figures above. Not a measurement, not investment advice.
- Editorial
a new entrant in SEO Research (1,073 products) needs a clear differentiator; competing on "also does seo research" will not be visible.
- Editorial
high concentration is a stronger barrier than size. In 139 segments one project already holds half the stars, so new projects there should integrate with the leader rather than replace it.
- Editorial
the 120 thin segments are where a single well-documented product can become the reference; check demand before building, because thin can also mean unwanted.
Sources
- Official MCP Registry: registry.modelcontextprotocol.io (external)checked
- GitHub topic search: docs.github.com (external)checked
- GitHub repository metadata: docs.github.com (external)checked
- AgentSidekick methodology: agentsidekick.dev (external)checked
Data behind this report
- Per-segment counts and launch momentum: /api/dataset/segments.jsonJSON, regenerated each build
- Taxonomy with keyword rules: /api/dataset/taxonomy.jsonJSON, regenerated each build
Cite this report
Most crowded categories in the agent ecosystem. AgentSidekick, 2026-11-21. https://www.agentsidekick.dev/reports/most-crowded-categories/
Figures change when the database is refreshed; cite the date shown.