---
fragment_peers:
  available-to-try: disponivel-para-experimentar
  experimental-with-known-limitations: experimental-com-limitacoes-conhecidas
  project-status: estado-do-projeto
  proposals-without-completed-implementations: propostas-ainda-sem-implementacao-concluida
  what-tests-need-to-investigate: o-que-os-testes-precisam-investigar
lang: en
peer: pt/status/
shared_fragments: false
title: Project status
---

# Project status

Legible is a draft, with formats, reference code and paths available for
initial tests that follow participants' work: what they can understand,
add and continue, and how much effort it takes. The economic hypothesis
needs to be examined through that use, beyond the technical correctness
of the files.

## Available to try

You can prepare a contribution with AI, review it and save it as a Legible Draft.
The first-use attachment contains instructions, the format and examples.
Start with [your first draft](first-draft.md). According to your intention, you
can [share and develop](participation.md), [participate in a space](get-started.md)
or [create one](open-space.md). The general guide includes an optional space model;
the participation guide is an alternative for that path. Choose one attachment.
The example space is fictional and optional; participants still configure access, channels and tools.
Participants exchange files they choose, preserving origins and relationships.

The site, book and paper are available in Portuguese and English. Portuguese
is also the complete normative source for draft 0.4 of the specification, the
NCO and NRC 0.2; English editions are synchronized reference translations. The
initial publication baseline uses v2 schemas.

## Experimental, with known limitations

The Rust, Python, MCP and local app components provide preparation, validation,
verification and local queries. The bench supports testing storage and access
per participant. Automated tests cover specific operations; use with actual
accounts, MCP clients, Android devices and networks requires checks in those
environments. Baseline signing uses plain Ed25519 and OpenSSH.

Versions identify different objects: specification draft 0.4, schemas
v2, Rust core 0.1.0, Python library and MCP server 0.0.1, and NRC 0.2. Updating software does not, on its own, change
record formats. Normative changes must identify their edition; existing records
retain their bytes and adopted convention.

## Proposals without completed implementations

The Python library delegates record building, sealing and verification to the Rust core. Its tests and the OpenSSH checks of fixtures provide evidence about specific operations; they do not by themselves demonstrate interoperability between two complete, independently developed implementations. That cross-test remains a maturity milestone to reach, without being presented as a requirement for trying this draft.

A specific Hermes integration, corporate package, contextual discovery across
multiple sources, native mobile signing and persistent identity recovery still
need development and testing. Consult the [implementation catalogue](implementations.md)
to distinguish available components from proposals.

## What tests need to investigate

Correct file exchange is a technical check. Assessing the thesis requires
finding out whether a contribution preserved before selection helped someone
ask another question, find an alternative or develop work they would not
have considered.

Count preparation and reading time, necessary corrections and each participant's
benefit too. Comparisons should provide equivalent access to people, models,
information and time. Failures and results without gains help identify where
the proposal needs to change.

[Create a space](open-space.md) · [Participate in a space](get-started.md) ·
[Submit a report or contribution to the protocol](contribute.md)