How to Use Actions
How to Use Actions
Choose the integration method that fits your architecture:
- MCP Server: AI assistants like Claude, Cursor, or Windsurf. Drop-in server that exposes actions as tools.
- A2A Protocol: Agent-to-agent communication. Google’s open protocol for autonomous systems.
- AI Toolset (TypeScript): Custom TypeScript agents. Full control with OpenAI, Vercel AI SDK, and more.
- AI Toolset (Python): Custom Python agents. Works with LangChain, CrewAI, PydanticAI, and more.
Test Actions
Test Actions
Try actions before integrating:
- Playground: Test actions in the StackOne dashboard with real data.
- Postman: Import our Postman collection to test the Actions RPC API directly.
- API Reference: Full API documentation for the Actions RPC endpoint.
How Actions Are Built
How Actions Are Built
StackOne creates actions by analyzing thousands of documented APIs using AI agents and manual QA. Each action is automatically tested, then manually verified by developers. Evaluation data continuously improves action quality.
Adding New Connectors
Adding New Connectors
Need a connector that’s not listed? Request it from StackOne (delivery depends on roadmap priorities unless on a plan with connector build SLAs) or build your own with Connector Engine.
Build a Connector with Connector Engine
Create custom connectors from any API documentation using AI-assisted generation.
All Connectors
Last Updated: 2026-01-23 · Data refreshed weekly via the /actions API endpoint. This list may include beta or preview connectors not visible by default in all organizations.