01 — Transparency
The record should survive.
Public protocol specifications and test data should remain public and discoverable, even after a feed is retired.
One model, every venue, every language, every definition, every database — compiled into any output (C, C++, C#, Lua, Python, Lean, Kaitai, DFDL, SQLite, kdb+, etc). Output is a pure function of the specification: identical on every run, and regenerated rather than maintained.
Financial organizations
Nasdaq, NYSE, CME, CBOE, LSEG, HKEX and more
Wire protocols documented
Market data and order entry — across 78 encoding formats
Generated output targets
7 languages · 6 definitions · 2 databases
Open source licensing
Least restrictive per repo · GPL-3.0 for Wireshark dissectors
Directory
A curated index of public protocol specifications, sample data, and reference material. These are collected from publicly available sources and preserved so the historical record survives even when exchanges retire or rename their feeds.
Principles
OMI is market-neutral. We do not take sides between exchanges, brokers, trading firms, or technology providers. We do hold positions on two things:
01 — Transparency
Public protocol specifications and test data should remain public and discoverable, even after a feed is retired.
02 — Best practices
Source code in the OMI ecosystem follows automated builds, regression testing, clean code, and SOLID principles.
Technology · Scaled Sources
The diagram below is a binary data model — a generic independent intermediate representation for binary protocols. Like any good compiler abstraction, it operates on an entire class of protocols at once, with no dependency between inputs and outputs. That independence removes the ceiling of a conventional code generator: N input languages times M output targets — a full M × N, defining a third-generation code generator.
Any input
Read and edit any binary specification in one representation — from ASN.1 and SBE to ITCH and beyond.
Any output
Represent any protocol, storage medium, or processing architecture — FPGA, Lua, C Wireshark dissectors, pcap to Parquet.
Common tooling
BPath and programmatic binary integration navigate, query, and manipulate any binary format — the foundation for advanced generation.
The result is a composable code-generation architecture — designed from the ground up with no limits. The compiler is developed by Scaled Sources and used to generate OMI repos.
Everything OMI produces is free, openly licensed, and community-driven. Browse the full catalog, contribute a protocol, or just start building.