HomeProjectsArticlesPK BoardAbout
Explore Projects

Klavis AI

Normal
GitHub Python Apache-2.0

Description

MCP integration platform that lets AI agents use tools reliably at any scale, providing MCP servers, clients, and integration solutions for production agent workflows.

Key Features

  • Strata intelligent connectors — optimize context windows for AI agents, smart tool call management
  • 100+ prebuilt integrations — out-of-the-box MCP integrations with OAuth support
  • MCP Sandbox — scalable MCP environments for LLM training and reinforcement learning
  • Multi-SDK support — Python SDK, TypeScript SDK, and REST API available
  • Self-hosting option — deploy via Docker or install Strata locally with pipx
  • Cloud-hosted service — use directly via klavis.ai without building infrastructure

Use Cases

💡 Agent tool integration — let AI agents reliably call external tools like Gmail, Slack
💡 MCP protocol standardization — provide unified MCP interfaces for different tools
💡 LLM training environment setup — build scalable training environments with MCP Sandbox
💡 Large-scale agent deployment — manage multiple MCP server instances in production via SDK
💡 Rapid prototype validation — use cloud-hosted service to quickly validate agent tool call approaches

Strengths & Limitations

✅ Strengths

  • • High community interest (5.8k stars)
  • • Permissive open-source license (Apache-2.0)

⚠️ Limitations

  • • High issue backlog (295 open issues)

Tags

mcp tool-integration mcp-server mcp-client python

Categories

🔌 MCP Tools

Quick Start

# Cloud-hosted (recommended)
# https://www.klavis.ai/docs/quickstart

# Self-host - Docker
docker pull ghcr.io/klavis-ai/github-mcp-server:latest
docker run -p 5000:5000 ghcr.io/klavis-ai/github-mcp-server:latest

# Self-host - Strata
pipx install strata-mcp
strata add --type stdio playwright npx @playwright/mcp@latest

# Python SDK
pip install klavis

from klavis import Klavis
from klavis.types import McpServerName

klavis = Klavis(api_key="your-key")
strata = klavis.mcp_server.create_strata_server(
    user_id="user123",
    servers=[McpServerName.GMAIL, McpServerName.SLACK]
)
Visit GitHub Visit Website

Related Projects

HexStrike AI

11.1k · Python
Active A

HexStrike AI is an advanced MCP server that lets AI agents autonomously run 150+ cybersecurity tools for automated pentesting, vulnerability discovery, and security research.

cybersecuritypentestingmcp-server +2
  • · 150+ professional security tools integrated — network recon, web app testing, password cracking, binary analysis, and cloud security
  • · 12+ autonomous AI agents for specialized tasks like BugBounty, CTF solving, CVE intelligence, and exploit generation
  • · MCP protocol integration with Claude, GPT, VS Code Copilot, Cursor, and other MCP-compatible clients
View full details → View on GitHub

Awesome Claude Skills

72.6k · Python
Active A

A curated list of awesome Claude Skills, resources and tools for customizing Claude AI workflows.

mcptoolspython +1
  • · 1000+ production-ready skills - Covering 9 categories including document processing, dev tools, data analysis, marketing, and writing
  • · Cross-agent compatible - Works with Claude Code, Claude.ai, OpenAI Codex, Cursor, Gemini CLI, Windsurf and more
  • · Progressive loading architecture - Each skill loads only name and description (~100 tokens), full content on demand, hundreds coexist
View full details → View on GitHub

MCP Context Forge

4.3k · Python
Active A

An AI Gateway, registry, and proxy by IBM that sits in front of any MCP, A2A, or REST/gRPC APIs, exposing a unified endpoint with centralized discovery, guardrails, and management.

mcpa2aapi-gateway +4
  • · Unified gateway federating MCP, A2A, and REST/gRPC APIs with a single endpoint and centralized governance
  • · Tools Gateway with MCP, REST, gRPC-to-MCP translation, and TOON compression support
  • · Agent Gateway supporting A2A protocol, OpenAI-compatible and Anthropic agent routing
View on GitHub

Related Articles

MCP安全OAuth

MCP Ecosystem Security: OAuth, Scope Isolation, and Auditing

MCP security design lags significantly behind feature development. A systematic deep dive into four threat models, OAuth 2.1 authorization, fine-grained scope design, path access control, sensitive data redaction, call auditing, MCP Gateway deployment, and tool supply chain security.

Read more →