AI Integrations¶
Agents can work with OrionBelt in two ways:
- Model the current questions. Author only the OBML needed for a task and submit
it with several independent queries to
POST /v1/oneshot/batch. The model can be temporary or retained for follow-up questions. The Agent Workflows guide includes a complete REST example that can be wrapped as a tool in your agent framework. - Query shared models. Discover established dimensions, measures, and metrics, then compose queries using the framework adapters below or the MCP server.
The ready-to-use REST integrations below use the shortcut endpoints and work best in admin-curated mode. Inline model authoring through the one-shot API requires a deployment that permits model uploads; admin-curated deployments accept batches referencing existing models.
Available Integrations¶
| Integration | Language | Description | Guide |
|---|---|---|---|
| LangChain / LangGraph | Python | 10 async tools + create_agent example |
integrations/langchain/ |
| OpenAI Agents SDK | Python | 10 function tools + Agent/Runner example |
integrations/openai-agents-sdk/ |
| CrewAI | Python | 10 tools + multi-agent crew example | integrations/crewai/ |
| Google ADK | Python | 10 FunctionTools + Gemini agent example | integrations/google-adk/ |
| Vercel AI SDK | TypeScript | 10 tools + Next.js API route example | integrations/vercel-ai-sdk/ |
| n8n | JSON | Workflow templates (batch + AI agent) | integrations/n8n/ |
| ChatGPT Custom GPT | OpenAPI | OpenAPI 3.1 spec + GPT instructions | integrations/chatgpt-custom-gpt/ |
| MCP Server | Python | 10+ tools for Claude, Copilot, Cursor, Windsurf | separate repo |
Each integration directory includes a self-contained README with setup instructions, code examples, and usage patterns. The MCP server is maintained as a separate repository.