OpenLine – a tiny receipt layer for agents (MCP/LangGraph ready)
OpenLine – a tiny receipt layer for agents (MCP/LangGraph ready)
• What: Given a small argument graph (nodes/edges), OpenLine returns a verifiable receipt (claim/because/but/so + guardrails + telemetry). • Why: Tracing is descriptive; receipts are normative. They make agent steps auditable & safe by default. • How: stdlib Python; MCP JSON-RPC stub; LangGraph node; schema + validator; optional signing step; GitHub Actions + Pages. Repo: https://github.com/terryncew/openline-core Demo: https://terryncew.github.io/COLE-Coherence-Layer-Engine-/nca... Would love feedback on the guard checks and the OLP schema—where should it grow?
Share cardActual performance
Launch Intel predictions
Analyze your own launch →Correct prediction on native model
Similar products
Breaklist – A Morning Briefing Printed on a Tiny Thermal Receipt
Linktree, but It's a Receipt
A tiny x86 compiler with ELF and PE target
A tiny WebAssembly compiler
Haskell Tiny Tetris
C++ (tiny) BigNum Multiply w/ Genaille-Lucas Rulers from 1891
TOCC – A Tiny DFA to C Compiler
Wrote a tiny WebAssembly (wat2wasm) compiler in Go
Tiny VPSs without exhausting IPv4 addresses
A tiny Go wrapper around Restcountries