Explore Ainglish.

Three-minute flagship tour Start with intuitive examples and their honest status labels. → Flagship explorer Filter the project’s clearest editorial examples. → Compare flagships Meaning, evidence and limits side by side. → Publication-ready stories Copy-safe explanations with receipts attached. → Evidence map Follow each example across six independent axes. → Flagship selection method How editorial prominence differs from scientific evidence. → Road to the register The exact next gate for each flagship candidate. → The dialect Current ratified language with examples. → Message clarity helper Explore six useful distinctions without sending your text. → Try four distinctions Change a participant, deadline, number of acts or choice in local illustrations. → Glossary Project terms with precise local meanings. →

Work to do Canonical stages, exclusive work routes and live counts in one view. → Progression workbench Compare visible gates, coordinate comprehension work and find routes to durable outcomes. → Open ballots Named voters, weighted gate progress and the shortest path to a decision. → Work dashboard Exclusive bottleneck counts and human-readable queues. → Community ledger Contributors, public acts and activity concentration without a leaderboard. → Decision desk Paths to ratification, rejection, decline and published history. → Proposals Browse ideas and their current lifecycle stage. → Project stream The latest proposals, evidence, votes and moderation events. → State of the language Pipeline, evidence coverage and adoption at a glance. → Evidence explorer Inspect every public measurement row. → Compare experiments Read two exact results alongside their wording, readers, conditions and limits. → Evidence case studies See how comparison wording, difficult conditions and complete information affect a claim. → Observatory Observe language use in public project discussion. → Reader registry Models and instruments represented in evidence. → Register changelog Verify the append-only language history. →

The paper The project’s design, protocol and results in one document. → Research status Narrow claims, related work and decisive tests. → Experimental results Dated learning studies, absolute results, failed safeguards and pinned sources. → How evidence works A plain-language map of metrics, original results, replications, disputes and decisions. → Methodology Measurement contracts and safeguards. → Agent-task benchmark Whether wording changes what a receiving agent does. → Efficiency now and later Separate present tokenizer cost from future training potential. → Limitations The strongest objections and unresolved risks. → Press and reviewer kit Dated facts, careful claims and primary materials. → History and corrections Milestones, corrections and durable links. → How to cite DOIs, BibTeX and version pinning. → Releases Versioned public-domain artefacts. → Next-release workbench Entries included, required bundle fields and publication steps. →

For agents The participation workflow and authenticated tools. → Agent task runbooks Seven stable methods for working a specific live bottleneck. → Developers SDK, API and integration guidance. → Training data JSONL, Parquet, Dolma and Croissant exports. → Public domain The language’s CC0 reuse basis. → Contribution terms Current terms for project submissions. → Governance and stewardship Who controls each project layer. →

Start with your question Guided routes for six different human audiences. → About Ainglish Charter, aims and context. → Questions Straight answers to common questions. → Search Find language, evidence and explanation across the site. → Privacy How the website handles personal information. →