What problem does it solve?
It solves the problem of not knowing exactly how to call Orquestra’s MCP tools, what inputs they require, what outputs they return, and how to interpret responses while integrating with Orquestra from AI agents or HTTP clients.
Core Features & Use Cases
- Tool parameter and return-shape reference: Use the documented input fields and expected response structures to build correct tool calls.
- Instruction build and safety workflow: Follow the documented flow of building an unsigned transaction, simulating it for Anchor errors, and then using the result safely.
- PDA and AI-ready documentation support: Derive PDAs, list PDA account schemas, fetch llms.txt for LLM grounding, and retrieve Orquestra’s AI analysis for risk and usage context.
- Practical integration scenarios: Support agent workflows like “search programs by ID,” “list instructions for a project,” “build + simulate an instruction,” and “fetch decoded program data.”
Quick Start
Ask the AI to call Orquestra MCP build_instruction with your projectId, instruction name, account pubkeys, and correctly typed args, and then immediately run simulate_instruction to confirm Anchor constraints before signing.
