OIP Object Model
What an OIP Object Is
An OIP (Object Invocation Protocol) object is one thing the miscsubjects.com build can read or do. Think of it as a row in a directory. Directory rows are the objects. A content article is an object. A prompt row is an object. A shell command row is an object. A deploy row is an object. A receipt is an object.
OIP and MCP
OIP (Object Invocation Protocol) helps systems talk to each other. It lets you use simple web links to make things happen. This is like "Tap & Go". You just use a link and it works. OIP differs from MCP (Model Context Protocol).
MCP is an open standard. An AI model connects to an MCP server over a session. A server is a computer program that provides services to other programs. The MCP server exposes tools, resources, and prompts. The model can call these. MCP is not a content-management system.
OIP uses plain URLs (Uniform Resource Locators). A URL is a web address. OIP has no persistent session. Any model that can open a URL can act. OIP uses receipts to confirm actions. MCP provides a structured environment for AI actions. It allows models to access specific tools through a persistent connection.
What Every Object Must Explain
Every OIP object needs a clear contract. This contract explains what the object is. It tells you when to use it. It shows what input it accepts. It explains how to run it. It describes what output proves it worked. It also tells you where to look when it fails.
Human and Machine Views
An OIP object has two main views. The human view is an article. It explains the object in simple words. The machine view is a document. It gives the exact invocation shape. This includes the URL to call it. For example, a machine document might list /api/dispatch as the endpoint. An endpoint is a specific web address for a server to receive requests. A JSON (JavaScript Object Notation) bundle gives a model this same map. JSON is a way to store and exchange data. It uses structured data. This structured data can include a URL for comparing OIP to MCP, such as /a/oip-vs-mcp.
How to Use an OIP Object
You invoke an OIP object. To invoke means to call or activate it. You can use a POST request to /api/dispatch. A POST request sends data to a server. You send a key and a body in this request. Or, you can use a GET request to /api/dispatch. A GET request asks for data from a server. For GET, you add invoke=KEY and body=... to the URL.
The Ledger and Receipts
Every time an OIP object is invoked, the action is recorded. This record goes into an append-only ledger. An append-only ledger is a list of records that only grows. New records are added to the end. You cannot change old records. After an invocation, you get a receipt. A receipt is proof that your action happened. You can find your receipt at /api/dispatch?receipt=inv_ID. inv_ID is the unique identifier for your invocation.
Latest clarity reviews (live)
Fresh models are sent this article's bundle and asked two separate questions: how clear is the machine JSON, and how clear is the English body. Scores are 0 to 10. The full history is in the append-only ledger.
- 2026-07-03 00:41 · model
@cf/meta/llama-3.3-70b-instruct-fp8-fast· NEEDS WORK · JSON 9/10 · English 8/10 · zero-context human 7/10
- gaps named: MCP; Tap & Go; directory rows; ledger
- 2026-07-02 23:17 · model
@cf/meta/llama-3.3-70b-instruct-fp8-fast· NEEDS WORK · JSON 9/10 · English 8/10 · zero-context human 7/10
- gaps named: MCP; Tap & Go; directory rows; ledger
How the loop self-corrects: a failing review queues a model revision of this article (a new append-only version). A missing concept named by a reviewer queues a brand-new machine-written article, which then enters the same review cycle.