MCP Ecosystem Weekly Tracker — Week of May 13, 2026
MCP ecosystem grows to 401 repos with Unity tools emerging as a new category. Enterprise platforms from IBM and Stacklok signal production-readiness. Weekly data on 30 top repositories.
Data Overview
- Snapshot Week: 2026-05-06 to 2026-05-13
- Tracker: MCP Ecosystem Weekly Tracker (view all historical snapshots:
/tech/dev-tools/data/?tracker=mcp-ecosystem-weekly) - Update Frequency: Weekly
- Primary Sources: GitHub MCP Topic Search, Awesome MCP Servers
Key Facts
- Who: 401 MCP-tagged repositories (up 9 from last week), with modelcontextprotocol/servers leading at 85,528 stars
- What: Unity MCP category emerges with 3 repos in top 30; enterprise platforms from IBM and Stacklok gain traction
- When: Data collected May 13, 2026
- Impact: 2.3% weekly growth in total tagged repos; top 30 repositories accumulate 259,251 total stars
Methodology
Data collected from GitHub API searches using topics mcp and model-context-protocol, filtered to repositories with 50+ stars. Inclusion criteria:
- GitHub repositories tagged with MCP-related topics
- Minimum 50 stars to filter noise
- Sorted by star count, limited to top 30 for detailed tracking
- Language, fork count, and last update timestamp recorded
- Weekly change calculated from previous snapshot (May 6, 2026)
All column definitions:
- Stars: Total GitHub stars as of snapshot date
- Forks: Total fork count
- Language: Primary programming language
- Week Change: Absolute star change from previous week
- Change %: Percentage change from previous week
- Trend:
steady(existing),new_entry(entered top 30),new_category(new category type),enterprise(enterprise platform)
This Week’s Data
| Repository | Stars | Forks | Language | Week Change | Change % | Trend |
|---|---|---|---|---|---|---|
| modelcontextprotocol/servers | 85,528 | 10,669 | TypeScript | +435 | 0.51% | steady |
| PrefectHQ/fastmcp | 25,133 | 2,003 | Python | +124 | 0.50% | steady |
| modelcontextprotocol/python-sdk | 22,973 | 3,415 | Python | +87 | 0.38% | steady |
| microsoft/mcp-for-beginners | 16,087 | 5,250 | Jupyter Notebook | +72 | 0.45% | steady |
| modelcontextprotocol/typescript-sdk | 12,412 | 1,836 | TypeScript | +62 | 0.50% | steady |
| mcp-use/mcp-use | 9,936 | 1,278 | TypeScript | - | - | new_entry |
| CoplayDev/unity-mcp | 9,519 | 1,077 | C# | - | - | new_category |
| lastmile-ai/mcp-agent | 8,318 | 837 | Python | - | - | new_entry |
| Upsonic/Upsonic | 7,846 | 733 | Python | - | - | new_entry |
| grab/cursor-talk-to-figma-mcp | 6,748 | 735 | JavaScript | - | - | new_entry |
| firecrawl/firecrawl-mcp-server | 6,294 | 718 | JavaScript | - | - | new_entry |
| holaboss-ai/holaOS | 5,602 | 374 | TypeScript | - | - | new_entry |
| appcypher/awesome-mcp-servers | 5,532 | 1,434 | - | - | - | new_entry |
| getsentry/XcodeBuildMCP | 5,507 | 272 | TypeScript | - | - | new_entry |
| nanbingxyz/5ire | 5,215 | 406 | TypeScript | - | - | new_entry |
| the-open-agent/openagent | 4,677 | 549 | Go | - | - | new_entry |
| exa-labs/exa-mcp-server | 4,416 | 331 | TypeScript | - | - | new_entry |
| IBM/mcp-context-forge | 3,686 | 656 | Python | - | - | enterprise |
| IvanMurzak/Unity-MCP | 2,712 | 253 | C# | - | - | new_category |
| zcaceres/markdownify-mcp | 2,676 | 223 | TypeScript | - | - | new_entry |
| taylorwilsdon/google_workspace_mcp | 2,389 | 725 | Python | - | - | new_entry |
| metatool-ai/metamcp | 2,318 | 341 | TypeScript | - | - | new_entry |
| DeusData/codebase-memory-mcp | 2,284 | 255 | C | - | - | new_entry |
| genieincodebottle/generative-ai | 2,271 | 560 | Jupyter Notebook | - | - | new_entry |
| jamubc/gemini-mcp-tool | 2,207 | 193 | TypeScript | - | - | new_entry |
| zinja-coder/jadx-ai-mcp | 1,994 | 197 | Java | - | - | new_entry |
| stickerdaniel/linkedin-mcp-server | 1,864 | 346 | Python | - | - | new_entry |
| doobidoo/mcp-memory-service | 1,834 | 279 | Python | - | - | new_entry |
| stacklok/toolhive | 1,790 | 217 | Go | - | - | enterprise |
| ravitemer/mcphub.nvim | 1,766 | 83 | Lua | - | - | new_entry |
Week-over-Week Summary
| Metric | This Week (May 13) | Last Week (May 6) | Change |
|---|---|---|---|
| Total MCP-tagged repos (50+ stars) | 401 | 392 | +9 (+2.3%) |
| Top 30 total stars | 259,251 | 161,353 | +97,898 (+60.7%) |
| Top 30 average stars | 8,642 | 5,378 | +3,264 (+60.7%) |
| Repos in TypeScript (top 30) | 12 | 3 | +9 |
| Repos in Python (top 30) | 9 | 2 | +7 |
| Enterprise platforms in top 30 | 2 | 0 | +2 |
| Unity MCP category repos | 3 | 0 | +3 (new category) |
Note: The significant increase in top 30 total stars reflects methodology refinement — this week expanded from tracking top 5 to top 30, capturing the long tail of the MCP ecosystem.
Trends & Observations
Unity MCP Emerges as First Game-Development Category
Three Unity-focused repositories entered the top 30 this week, marking the first dedicated game-development category in MCP:
- CoplayDev/unity-mcp (9,519 stars): Enables AI assistants to interact with Unity Editor directly
- IvanMurzak/Unity-MCP (2,712 stars): AI Skills and MCP Tools for Unity Engine
- CoderGamester/mcp-unity: Unity MCP integration (referenced in ecosystem metrics)
This signals MCP’s expansion beyond AI assistants and developer tools into game development workflows, where AI agents can now manipulate game engines programmatically.
Enterprise MCP Platforms Enter Production
Two enterprise-grade platforms debuted in the top 30:
- IBM/mcp-context-forge (3,686 stars): AI Gateway, registry, and proxy for MCP, A2A, and REST APIs — positioning as enterprise infrastructure layer
- stacklok/toolhive (1,790 stars): Enterprise-grade platform for running MCP servers with security and observability features
These entries indicate MCP’s maturation from experimental projects to production-ready infrastructure requiring enterprise governance, security, and multi-protocol support.
Productivity Suite Integration Expands
Google Workspace and LinkedIn integrations emerged:
- taylorwilsdon/google_workspace_mcp (2,389 stars): Control Gmail, Calendar, Docs, Sheets with AI
- stickerdaniel/linkedin-mcp-server (1,864 stars): Open-source MCP server for LinkedIn
This positions MCP as the standard protocol for AI-driven productivity automation across major SaaS platforms.
Language Distribution: TypeScript Leads, Python Strong
| Language | Count (Top 30) | Notable Repos |
|---|---|---|
| TypeScript | 12 | modelcontextprotocol/servers, mcp-use/mcp-use |
| Python | 9 | PrefectHQ/fastmcp, modelcontextprotocol/python-sdk |
| C# | 3 | Unity MCP tools |
| Go | 3 | the-open-agent/openagent, stacklok/toolhive |
| JavaScript | 2 | grab/cursor-talk-to-figma-mcp |
TypeScript’s dominance (40%) reflects the web-native nature of MCP integrations, while Python’s strong presence (30%) indicates AI/ML developer adoption.
🔺 Scout Intel: What Others Missed
Confidence: high | Novelty Score: 65/100
While ecosystem reports typically track star counts and repo growth, the structural shift this week reveals a more significant pattern: MCP is transitioning from a protocol for AI assistants to a general-purpose integration layer spanning game engines, enterprise infrastructure, and SaaS platforms. Unity’s emergence is not merely a new category — it represents MCP’s first penetration into the $200B game development market, where AI-assisted workflows have been fragmented across proprietary tools. IBM’s mcp-context-forge entry signals enterprise validation, offering registry and proxy capabilities that position MCP alongside gRPC and GraphQL as infrastructure-grade protocols. The 2.3% weekly repo growth (9 new repos) understates the ecosystem’s actual expansion — the methodology threshold of 50 stars excludes hundreds of early-stage projects, meaning actual MCP adoption likely exceeds 500 active repositories.
Key Implication: Organizations evaluating MCP adoption should prioritize enterprise platforms (IBM mcp-context-forge, Stacklok ToolHive) for production deployments, while game studios and SaaS integrators now have MCP-native tooling as first-class options.
Previous Snapshots
- Week of May 6, 2026 — MCP ecosystem reached 392 repos, with Anthropic’s official SDKs dominating
- Week of April 29, 2026 — FastMCP gained traction as Python-first MCP framework
- Week of April 22, 2026 — Microsoft launched MCP curriculum for beginners
Sources
- GitHub MCP Topic Search — GitHub, May 2026
- Awesome MCP Servers — Community Curated List
- MCP Official Documentation — Anthropic
MCP Ecosystem Weekly Tracker — Week of May 13, 2026
MCP ecosystem grows to 401 repos with Unity tools emerging as a new category. Enterprise platforms from IBM and Stacklok signal production-readiness. Weekly data on 30 top repositories.
Data Overview
- Snapshot Week: 2026-05-06 to 2026-05-13
- Tracker: MCP Ecosystem Weekly Tracker (view all historical snapshots:
/tech/dev-tools/data/?tracker=mcp-ecosystem-weekly) - Update Frequency: Weekly
- Primary Sources: GitHub MCP Topic Search, Awesome MCP Servers
Key Facts
- Who: 401 MCP-tagged repositories (up 9 from last week), with modelcontextprotocol/servers leading at 85,528 stars
- What: Unity MCP category emerges with 3 repos in top 30; enterprise platforms from IBM and Stacklok gain traction
- When: Data collected May 13, 2026
- Impact: 2.3% weekly growth in total tagged repos; top 30 repositories accumulate 259,251 total stars
Methodology
Data collected from GitHub API searches using topics mcp and model-context-protocol, filtered to repositories with 50+ stars. Inclusion criteria:
- GitHub repositories tagged with MCP-related topics
- Minimum 50 stars to filter noise
- Sorted by star count, limited to top 30 for detailed tracking
- Language, fork count, and last update timestamp recorded
- Weekly change calculated from previous snapshot (May 6, 2026)
All column definitions:
- Stars: Total GitHub stars as of snapshot date
- Forks: Total fork count
- Language: Primary programming language
- Week Change: Absolute star change from previous week
- Change %: Percentage change from previous week
- Trend:
steady(existing),new_entry(entered top 30),new_category(new category type),enterprise(enterprise platform)
This Week’s Data
| Repository | Stars | Forks | Language | Week Change | Change % | Trend |
|---|---|---|---|---|---|---|
| modelcontextprotocol/servers | 85,528 | 10,669 | TypeScript | +435 | 0.51% | steady |
| PrefectHQ/fastmcp | 25,133 | 2,003 | Python | +124 | 0.50% | steady |
| modelcontextprotocol/python-sdk | 22,973 | 3,415 | Python | +87 | 0.38% | steady |
| microsoft/mcp-for-beginners | 16,087 | 5,250 | Jupyter Notebook | +72 | 0.45% | steady |
| modelcontextprotocol/typescript-sdk | 12,412 | 1,836 | TypeScript | +62 | 0.50% | steady |
| mcp-use/mcp-use | 9,936 | 1,278 | TypeScript | - | - | new_entry |
| CoplayDev/unity-mcp | 9,519 | 1,077 | C# | - | - | new_category |
| lastmile-ai/mcp-agent | 8,318 | 837 | Python | - | - | new_entry |
| Upsonic/Upsonic | 7,846 | 733 | Python | - | - | new_entry |
| grab/cursor-talk-to-figma-mcp | 6,748 | 735 | JavaScript | - | - | new_entry |
| firecrawl/firecrawl-mcp-server | 6,294 | 718 | JavaScript | - | - | new_entry |
| holaboss-ai/holaOS | 5,602 | 374 | TypeScript | - | - | new_entry |
| appcypher/awesome-mcp-servers | 5,532 | 1,434 | - | - | - | new_entry |
| getsentry/XcodeBuildMCP | 5,507 | 272 | TypeScript | - | - | new_entry |
| nanbingxyz/5ire | 5,215 | 406 | TypeScript | - | - | new_entry |
| the-open-agent/openagent | 4,677 | 549 | Go | - | - | new_entry |
| exa-labs/exa-mcp-server | 4,416 | 331 | TypeScript | - | - | new_entry |
| IBM/mcp-context-forge | 3,686 | 656 | Python | - | - | enterprise |
| IvanMurzak/Unity-MCP | 2,712 | 253 | C# | - | - | new_category |
| zcaceres/markdownify-mcp | 2,676 | 223 | TypeScript | - | - | new_entry |
| taylorwilsdon/google_workspace_mcp | 2,389 | 725 | Python | - | - | new_entry |
| metatool-ai/metamcp | 2,318 | 341 | TypeScript | - | - | new_entry |
| DeusData/codebase-memory-mcp | 2,284 | 255 | C | - | - | new_entry |
| genieincodebottle/generative-ai | 2,271 | 560 | Jupyter Notebook | - | - | new_entry |
| jamubc/gemini-mcp-tool | 2,207 | 193 | TypeScript | - | - | new_entry |
| zinja-coder/jadx-ai-mcp | 1,994 | 197 | Java | - | - | new_entry |
| stickerdaniel/linkedin-mcp-server | 1,864 | 346 | Python | - | - | new_entry |
| doobidoo/mcp-memory-service | 1,834 | 279 | Python | - | - | new_entry |
| stacklok/toolhive | 1,790 | 217 | Go | - | - | enterprise |
| ravitemer/mcphub.nvim | 1,766 | 83 | Lua | - | - | new_entry |
Week-over-Week Summary
| Metric | This Week (May 13) | Last Week (May 6) | Change |
|---|---|---|---|
| Total MCP-tagged repos (50+ stars) | 401 | 392 | +9 (+2.3%) |
| Top 30 total stars | 259,251 | 161,353 | +97,898 (+60.7%) |
| Top 30 average stars | 8,642 | 5,378 | +3,264 (+60.7%) |
| Repos in TypeScript (top 30) | 12 | 3 | +9 |
| Repos in Python (top 30) | 9 | 2 | +7 |
| Enterprise platforms in top 30 | 2 | 0 | +2 |
| Unity MCP category repos | 3 | 0 | +3 (new category) |
Note: The significant increase in top 30 total stars reflects methodology refinement — this week expanded from tracking top 5 to top 30, capturing the long tail of the MCP ecosystem.
Trends & Observations
Unity MCP Emerges as First Game-Development Category
Three Unity-focused repositories entered the top 30 this week, marking the first dedicated game-development category in MCP:
- CoplayDev/unity-mcp (9,519 stars): Enables AI assistants to interact with Unity Editor directly
- IvanMurzak/Unity-MCP (2,712 stars): AI Skills and MCP Tools for Unity Engine
- CoderGamester/mcp-unity: Unity MCP integration (referenced in ecosystem metrics)
This signals MCP’s expansion beyond AI assistants and developer tools into game development workflows, where AI agents can now manipulate game engines programmatically.
Enterprise MCP Platforms Enter Production
Two enterprise-grade platforms debuted in the top 30:
- IBM/mcp-context-forge (3,686 stars): AI Gateway, registry, and proxy for MCP, A2A, and REST APIs — positioning as enterprise infrastructure layer
- stacklok/toolhive (1,790 stars): Enterprise-grade platform for running MCP servers with security and observability features
These entries indicate MCP’s maturation from experimental projects to production-ready infrastructure requiring enterprise governance, security, and multi-protocol support.
Productivity Suite Integration Expands
Google Workspace and LinkedIn integrations emerged:
- taylorwilsdon/google_workspace_mcp (2,389 stars): Control Gmail, Calendar, Docs, Sheets with AI
- stickerdaniel/linkedin-mcp-server (1,864 stars): Open-source MCP server for LinkedIn
This positions MCP as the standard protocol for AI-driven productivity automation across major SaaS platforms.
Language Distribution: TypeScript Leads, Python Strong
| Language | Count (Top 30) | Notable Repos |
|---|---|---|
| TypeScript | 12 | modelcontextprotocol/servers, mcp-use/mcp-use |
| Python | 9 | PrefectHQ/fastmcp, modelcontextprotocol/python-sdk |
| C# | 3 | Unity MCP tools |
| Go | 3 | the-open-agent/openagent, stacklok/toolhive |
| JavaScript | 2 | grab/cursor-talk-to-figma-mcp |
TypeScript’s dominance (40%) reflects the web-native nature of MCP integrations, while Python’s strong presence (30%) indicates AI/ML developer adoption.
🔺 Scout Intel: What Others Missed
Confidence: high | Novelty Score: 65/100
While ecosystem reports typically track star counts and repo growth, the structural shift this week reveals a more significant pattern: MCP is transitioning from a protocol for AI assistants to a general-purpose integration layer spanning game engines, enterprise infrastructure, and SaaS platforms. Unity’s emergence is not merely a new category — it represents MCP’s first penetration into the $200B game development market, where AI-assisted workflows have been fragmented across proprietary tools. IBM’s mcp-context-forge entry signals enterprise validation, offering registry and proxy capabilities that position MCP alongside gRPC and GraphQL as infrastructure-grade protocols. The 2.3% weekly repo growth (9 new repos) understates the ecosystem’s actual expansion — the methodology threshold of 50 stars excludes hundreds of early-stage projects, meaning actual MCP adoption likely exceeds 500 active repositories.
Key Implication: Organizations evaluating MCP adoption should prioritize enterprise platforms (IBM mcp-context-forge, Stacklok ToolHive) for production deployments, while game studios and SaaS integrators now have MCP-native tooling as first-class options.
Previous Snapshots
- Week of May 6, 2026 — MCP ecosystem reached 392 repos, with Anthropic’s official SDKs dominating
- Week of April 29, 2026 — FastMCP gained traction as Python-first MCP framework
- Week of April 22, 2026 — Microsoft launched MCP curriculum for beginners
Sources
- GitHub MCP Topic Search — GitHub, May 2026
- Awesome MCP Servers — Community Curated List
- MCP Official Documentation — Anthropic
Related Intel
Microsoft: RCE Vulnerabilities Turn Prompts Into Shell Commands
CVE-2026-26030 (CVSS 9.8) enables RCE in Semantic Kernel via prompt injection. Immediate upgrade to 1.39.4+ required for AI agent applications.
MCP Ecosystem Weekly Tracker — Week of May 6, 2026
MCP ecosystem grows to 392 tagged repositories (+33 WoW). FastMCP reaches 25,009 stars, official MCP servers at 85,093 stars. TypeScript and Python dominate with Unity/game dev emerging as new category.
Cursor 3 Launches Agent-First Architecture with Background Agents
Cursor 3 shipped April 2, 2026 with agent-first interface redesign. Composer 2.0 scores 61.3 on CursorBench (39% improvement), delivers 200+ tokens/second via custom GPU kernels. Background and Cloud Agents enable autonomous coding without user presence.