Open source MCP Servers for AWS
9.4k
Stars
1.6k
Forks
833
Open issues
30
Contributors
AI Analysis
This is a suite of specialized MCP (Model Context Protocol) servers that enable AI agents and LLMs to interact with AWS services. It serves developers and organizations building AI applications that need programmatic access to AWS infrastructure, from compute and databases to cost management and documentation. The project is specifically designed for integration with MCP-compatible clients and is not a general-purpose AWS SDK.
Inferred from signals mentioned in the README (tests, CI, type safety) — not a review of the actual code.
AI's overall editorial judgment — not an average of the bars above, can weigh other factors too.
AWS builds official MCP servers to wire AI agents directly into its cloud services
awslabs/mcp is a suite of Model Context Protocol servers maintained by AWS that expose AWS services — documentation, CDK, Lambda, ECS, Bedrock, DynamoDB, S3, Cost Explorer, and more — as tools callable by LLM-based agents and coding assistants. It is built for developers using AI coding environments (Cursor, Claude Code, VS Code, Windsurf, Cline, Kiro) who want agents to query, deploy, and operate AWS infrastructure. With 9,300+ stars and active pushes as of June 2026, it is among the most adopted AWS-specific MCP collections. AWS has since announced a successor product (Agent Toolkit for AWS), which signals strategic maturation rather than abandonment of this repo.
Created March 2025, coinciding with the broader MCP ecosystem surge after Anthropic released the Model Context Protocol spec. Grew rapidly as the canonical AWS-blessed MCP integration hub.
Initial growth was driven by the MCP ecosystem's explosive adoption in early-to-mid 2025 and AWS's official backing lending credibility. Stars plateaued somewhat (29 in last 7 days as of June 2026), likely because AWS has redirected new users toward Agent Toolkit for AWS, its production-grade successor announced May 2026.
AWS's own announcement of Agent Toolkit for AWS cites feedback from customers using these MCP servers, implying real user base. The 1,593 forks and 9,323 stars for a repo less than 16 months old in a specialized category suggest meaningful adoption. Multiple supported IDE integrations (Cursor, Claude Code, VS Code, Windsurf, Cline, Kiro) point to practical use. However, no third-party usage studies or download metrics are cited in the README.
Appears to be a monorepo of independent Python-based MCP servers, each targeting a specific AWS service or domain. Likely uses the official MCP SDK for Python. README mentions both local (stdio) and remote (HTTP/SSE, with SSE being removed) transport mechanisms, and an 'MCP AWS Lambda Handler Module' suggesting Lambda-deployable remote server support. Individual servers appear installable via uvx/pip separately.
Codecov badge is present in the README, indicating automated coverage tracking exists. Exact coverage percentage is not stated in the README excerpt but the badge implies it is being measured and likely non-trivial for an AWS Labs project.
Last push was June 23, 2026 — the same day as this analysis — indicating active, current maintenance. The repo explicitly states it continues to accept contributions and that useful projects will migrate to Agent Toolkit for AWS over time. OSSF Scorecard badge present, reflecting security hygiene awareness. Maintenance appears healthy.
ADOPT IF: you are building AI agent workflows on AWS today and want the fastest, officially-supported path to exposing AWS services as agent tools in coding assistants — this is the reference implementation. AVOID IF: you need production-grade IAM auditability, CloudTrail integration, or hardened agentic permissions — the newer Agent Toolkit for AWS is explicitly recommended for those cases. MONITOR IF: you are evaluating which project to standardize on long-term; AWS's stated direction is to migrate the most valuable servers into Agent Toolkit, so the boundary between the two products will shift over the next 12-18 months.
Independent dimensions
Mainstream potential
5/10
Technical importance
7/10
Adoption evidence
6/10
- AWS has publicly positioned Agent Toolkit for AWS as the production successor, which may reduce investment in this repo over time and cause contributors to migrate.
- The MCP protocol itself is still evolving; breaking changes in the spec (e.g., SSE transport removal already happened) may require frequent updates across many servers.
- A monorepo covering dozens of AWS services means quality and maintenance depth likely varies significantly between servers — some may lag behind AWS API changes.
- Security surface area is broad: agents using these servers can perform real AWS operations, and misconfigured IAM permissions could lead to unintended resource changes or cost overruns.
- Adoption plateau risk: the 29 stars/week figure (June 2026) is modest, suggesting the repo may be entering a maintenance phase rather than a growth phase as the ecosystem consolidates around Agent Toolkit.
This repo will remain a useful open reference and community experimentation ground, but net-new feature development will increasingly happen in Agent Toolkit for AWS. Expect gradual migration of mature servers out of this repo over 12-24 months.
Newsletter
Get analyses like this every Monday
Free weekly digest of the most interesting open-source discoveries.
Languages
Information
- Website
- https://awslabs.github.io/mcp/
- Language
- Python
- License
- Apache-2.0
- Last updated
- 12h ago
- Created
- 16mo ago
- Analyzed with
- anthropic/claude-haiku-4-5
Stars over time
Contributors over time
Top 100 contributors only — repos with more will plateau at 100.
Open issues
aws-api-mcp-server: migration to AWS MCP Server
(module name): (short issue description)
Multi-profile support
[mysql-mcp-server] Add --secret_arn support for custom Secrets Manager credentials (parity with mssql-mcp-server)
Open pull requests
feat(lambda): adding mcp server prompts support to MCPLambdaHandler
feat(cloudwatch-appsignals-mcp-server): Node.js docker sample app
feat: Add comprehensive PCAP Analyzer MCP Server (31 network analysis tools) with fixes in test cases and precommit
Top contributors
Similar repos
aws/agent-toolkit-for-aws
Agent Toolkit for AWS provides MCP servers, skills, and plugins that enable AI...
appcypher/awesome-mcp-servers
This is a curated directory of Model Context Protocol (MCP) servers that enable...
lastmile-ai/mcp-agent
mcp-agent is a Python framework for building AI agents using the Model Context...
YuzeHao2023/Awesome-MCP-Servers
This is a curated community-driven list of Model Context Protocol (MCP)...
| Repository | Stars | Week Δ | Language | Score | Updated |
|---|---|---|---|---|---|
|
awslabs/mcp
this
|
9.4k | +36 | Python | 7/10 | 12h ago |
|
aws/agent-toolkit-for-aws
|
1.8k | — | Python | 8/10 | 9h ago |
|
appcypher/awesome-mcp-servers
|
5.7k | — | — | 7/10 | 2mo ago |
|
lastmile-ai/mcp-agent
|
8.4k | — | Python | 8/10 | 6mo ago |
 microsoft/mcp
|
3.4k | — | C# | 7/10 | 2d ago |
|
YuzeHao2023/Awesome-MCP-Servers
|
1.1k | — | — | 7/10 | 2mo ago |
Microsoft's equivalent for Azure DevOps, scoped narrower (one service vs. AWS's multi-service suite). 1,838 stars vs. 9,323, suggesting AWS's offering has broader traction, possibly because AWS has more services to expose and a larger developer agent ecosystem.
A single-vendor MCP integration (Atlassian), community-driven with 5,439 stars. Comparable niche focus but lacks official vendor backing. awslabs/mcp benefits from direct AWS team involvement and IAM/auth alignment.
AWS's own production-grade successor with IAM condition keys, CloudWatch/CloudTrail visibility, and accuracy-evaluated skills. This repo remains the open, community-hackable option; the Agent Toolkit targets production deployments. Both coexist for now.
A community directory listing MCP servers including AWS-related ones. Not a direct technical competitor — it indexes projects like awslabs/mcp rather than replacing them. Signals how large the broader MCP ecosystem has grown (89,644 stars).
An orchestration framework for MCP agents rather than a collection of AWS service integrations. Complementary rather than competing — one could use mcp-agent to orchestrate awslabs/mcp servers.
