Self-extending agents with enterprise-grade control
137+ production tools. 10 meta-tools that let agents create, validate, and approve new capabilities at runtime. Policy engine with risk classification. Human-in-the-Loop approval for critical actions. One YAML definition works across TypeScript, Python, LangChain, CrewAI, Claude MCP.
🔧 Meta-Tools let agents create new capabilities mid-conversation. 🛡️ Policy Engine classifies risk and enforces rules. 🤝 HITL requires human approval for critical actions. Other SDKs give agents a toolbox. Matimo gives them a workshop — with safety guardrails.
import { MatimoInstance } from 'matimo';
const matimo = await MatimoInstance.init();
const result = await matimo.execute('slack_send_channel_message', {
channel: '#general',
text: 'Hello Matimo!'
});Matimo doesn't compete—it complements. We work alongside your existing frameworks.
Whether you're using LangChain for orchestration, CrewAI for multi-agent teams, or Claude for intelligence, Matimo adds tools + skills + policy engine that enhances what you already have. One definition in YAML. Deploy everywhere.
| Feature | |
 |
 |
|---|---|---|---|
| Skills System (> just tools) | ✅ Yes | ❌ No | ❌ No |
| Policy Engine (approval, risk gates) | ✅ Built-in | ❌ No | ❌ No |
| TypeScript SDK | ✅ Yes | ✅ Yes | ⚠️ Limited |
| Python SDK (LangChain, CrewAI compatible) | ✅ Yes | ✅ Yes | ✅ Yes |
| Works with ANY framework | ✅ Yes | ✅ Yes | Purpose-built |
| MCP Server (Claude Desktop, Cursor) | ✅ Yes | ❌ No | ❌ No |
| Agent Self-Extension (create tools at runtime) | ✅ Yes | ❌ No | ❌ No |
| YAML-first (single definition) | ✅ Yes | JSON schema | Python decorators |
| Semantic Skill Search (TF-IDF) | ✅ Yes | ❌ No | ❌ No |
| Open Source | ✅ MIT | ✅ MIT | ✅ Apache 2.0 |
| Pre-Built Tools (Slack, GitHub, Gmail, etc.) | ✅ Growing | ✅ Extensible | ✅ Extensible |
Add a skills layer + policy engine to your existing LangChain agents. NO refactoring needed.
Agents that learn from past runs. Semantic skill discovery + policy-gated approval. Multi-crew orchestration.
Self-extending agents that build new tools at runtime. Policy engine keeps them safe + audited.
Deploy to any framework or agent runtime with zero code changes
Convert tools to `StructuredTool` instantly. Works with ReAct, OpenAI Functions, and any LangChain agent executor.
Seamless `BaseTool` conversion with async support. Multi-agent crews and hierarchical workflows out of the box.
Run Matimo as an MCP server. Claude Desktop, Cursor, Windsurf instantly see all your tools. One command: `npx matimo mcp`.
Native SDK for custom agent loops. Factory pattern, decorator pattern, or direct SDK access—build however you want.
Tools via command-line, HTTP middleware, or direct invocation. Integrate into any workflow or service.
OpenAI, Anthropic, Google, Ollama—Matimo is agnostic. Your tools work with any LLM or agent framework.
Everything you need to build AI agents with confidence
Official Python 3.11+ support with async/await. Identical APIs, full feature parity, >95% test coverage. Works with LangChain, CrewAI, MCP.
10 meta-tools let agents autonomously create, validate, approve, and hot-reload new tools at runtime — without human intervention for trusted operations.
Agents learn from built-in SKILL.md files via TF-IDF semantic search. Progressive disclosure loads only relevant context — keeping agents fast and token-efficient.
YAML-driven approval tiers: auto, approval-required, or blocked. HITL quarantine, content validation, and integrity tracking keep agent-created tools safe.
Dual-transport MCP server (stdio + HTTP). One command to start: npx matimo mcp. One command to configure Claude Desktop: npx matimo mcp setup. Tools auto-discovered, secrets injected server-side, zero config required.
Write tools in simple YAML. No duplicating schemas across frameworks. One definition, infinite possibilities.
Plug into LangChain, OpenAI agents, Claude, custom decorators—no framework lock-in or re-implementation.
Start with Slack, Gmail, GitHub, Notion, Hubspot, Postgres tools. Add your own in minutes.
Provider-agnostic OAuth2 support. Secure auth for Google, GitHub, Slack, and any service.
Full TypeScript support with runtime validation. Zod-powered parameter and response schemas.
Works with any AI framework. LangChain, CrewAI, custom agents—same tools, different frameworks.
Everything you need to build production-grade AI agents
Works with any LLM framework. Deploy once, use everywhere.
Define tools once in YAML, execute across TypeScript, Python, and any framework.
Update tools at runtime without restarting your application.
Zod (TypeScript) & Pydantic (Python) validation for all inputs and outputs.
Send messages, create channels, manage conversations and threads.
Create issues/PRs, update repos, search code, manage workflows.
Send emails, manage messages, create drafts with full templating.
Create pages, query databases, manage content at scale.
Execute SQL queries safely with full parameterization.
Manage contacts, deals, companies, and customer data.
Create campaigns, manage subscribers and audiences at scale.
Send SMS/MMS messages, manage phone communications globally.
Automatic risk assessment: low, medium, high, critical with customizable policies.
Human-in-the-loop workflow management. Require approval for sensitive operations.
Complete event logging for compliance and debugging. Never miss what agents do.
Pattern-based security rules. Prevent injection attacks and harmful content.
Hash-based tamper detection. Verify tools haven't been modified unexpectedly.
Time-based approval expiry. Approvals expire automatically for security.
TF-IDF powered skill discovery. Find the right capability by natural language.
Markdown-based skills with YAML frontmatter. Context injection for agents.
10 built-in tools that create, validate, and approve new capabilities at runtime.
Agents create tools dynamically. Your agents learn and grow during execution.
Convert tools to StructuredTool. Add Matimo to existing LangChain agents instantly.
Full CrewAI integration. Convert to BaseTool. Multi-crew orchestration support.
Run as standalone MCP. Claude Desktop, Cursor, and any MCP client see all tools.
Provider-agnostic OAuth2 flows. Secure token management and refresh.
Environment vars, .env files, Vault, AWS Secrets Manager. Resolver chain pattern.
Secrets never logged. Automatic redaction in error messages and debug output.
Comprehensive test coverage. 97% coverage across TypeScript and Python.
Winston (TS) & Stdlib logging (Python). JSON format for log aggregation.
Exponential backoff, configurable retry policies for resilience.
Comprehensive error codes, rich context, fail-closed security model.
Choose what works best for your use case. Both TypeScript and Python SDKs support all patterns.
Simple and direct. Load tools and execute by name. Perfect for CLI tools and quick prototypes.
Python
Slack
Gmail
GitHub
Notion
PostgreSQL
HubSpot
Bruno
Mailchimp
Twilio
const matimo = await MatimoInstance.init();
await matimo.execute('tool-name', params);Class-based approach. Use @tool decorator for natural, object-oriented tool execution.
@tool('tool-name')
async myMethod(param) {
// Auto-executed
}Full LangChain, CrewAI integration. Convert tools to framework schemas automatically.
const tools = matimo
.listTools()
.map(t => toLangChainSchema(t));Run Matimo as a standalone MCP server. Claude Desktop, Cursor, and any MCP client see all your tools instantly.
// stdio (Claude Desktop)
npx matimo mcp
// HTTP (remote agents)
npx matimo mcp --transport httpFirst-class CLI for environment setup, validation, and approval workflows. Works with both TypeScript and Python SDKs.
Node: npx matimo doctor
Python: matimo doctor
Diagnoses your setup: versions, @matimo/* packages, missing env variables, and YAML validation.
Node: npx matimo review list
Python: matimo review list
Lists all tools pending approval or in quarantine. Essential for HITL (human-in-the-loop) governance.
Node: npx matimo mcp setup
Python: matimo mcp setup
Auto-generates and configures Claude Desktop integration. One command, instant setup.
Everything you need to get up and running in minutes
5-minute walkthrough. Set up your first tool in minutes with step-by-step instructions for both TypeScript and Python.
YAML-first design philosophy. Build HTTP, command, and function-based tools with built-in validation and error handling.
40+ runnable examples. LangChain, CrewAI, MCP, native patterns—all tested and production-ready.
Complete TypeScript and Python API docs. Every class, method, and parameter documented with examples.
Domain knowledge layer for agents. Learn how to build, discover, and leverage skills for context-aware execution.
Get help from maintainers and community. Ask questions, share tools, and stay updated with the latest releases.
Enterprise-grade testing, security, and observability baked in
2,541 tests passing (TypeScript + Python). Comprehensive unit, integration, and security test suites. Industry-leading quality standards.
6 critical security patches. Secret isolation, command injection prevention, integrity tracking, and real-time threat detection via policy engine.
10 provider packages with 110+ production-ready tools. Slack, GitHub, Gmail, Notion, HubSpot, Mailchimp, Postgres, Twilio, and growing.
YAML-driven governance with auto/approval-required/blocked tiers. HITL quarantine, approval manifests, and audit trails for compliance.
Shared executor pools, efficient TF-IDF embeddings, hot-reload atomicity. Built for production scale with CrewAI, LangChain, and beyond.
Identical APIs across TypeScript and Python. Cross-language parity means teams can ship together. One YAML definition, infinite deployments.
Connect with developers, share ideas, and get help building with Matimo
Building with us. Ask questions, share your tools, and stay updated.
Join the growing community building with Matimo. Start in minutes.