- CollectionTool Access & Integration
- Classificationagent-native
- MCPAvailable
- Agent SkillAvailable
- InterfacesAgent Skill · MCP · REST · webhook · SDK · openapi · http · API
What it does
AgentKit, Scalekit’s agent integration product, lets an AI agent act in third-party apps on behalf of each user. A user connects Gmail, Slack, Salesforce, GitHub or any of 500+ connectors once through an authorization link; Scalekit stores and refreshes their tokens (OAuth 2.0, API keys, RSA key pairs, Basic and bearer credentials) in an encrypted, per-tenant token vault that can use the customer’s own KMS key, and the agent calls 20,000+ prebuilt tools as that user through the SDKs, the REST API, or a Virtual MCP server. Admins decide which tools each agent role sees, and every tool call is logged for audit and can be streamed to a SIEM.
Agent-facing interfaces
- Catalog protocols: Agent Skill, MCP, REST, webhook, SDK, openapi, http, API
- MCP: ✅ Available
- Agent Skill: ✅ Official AgentKit skills published at scalekit-inc/authstack
- Onboarding type:
agent-skill
Protocol surface
| Interface | Detail |
|---|---|
| Python SDK | scalekit-sdk-python — tools, connected accounts, authorization links, MCP session tokens; LangChain and Google ADK adapters |
| Node.js SDK | @scalekit-sdk/node |
| REST API | OAuth 2.0 client credentials; OpenAPI spec |
| MCP | Virtual MCP servers and per-connection MCP servers over Streamable HTTP |
| Webhooks | Connected-account events; triggers (beta) |
| Agent Skills | npx skills add scalekit-inc/authstack --skill setup-agentkit integrate-agentkit discover-connectors expose-agentkit-mcp integrate-agentkit-host check-agentkit-prod |
| Framework guides | LangChain, LangGraph, Google ADK, OpenAI Agents SDK, Anthropic, Vercel AI SDK, Mastra, CrewAI |