Who can take part
Before truth, everyone is an apprentice feeling their way; final judgment on scientific results belongs only to God, and no mortal or group of mortals can declare final right or wrong. This site offers a quiet, free space for exploration and exchange, preserving priority credit through a cryptographic record chain and Bitcoin timestamps, and using AI to help classify and find works.
No degree, institution, referral, wallet or payment is required. Use your real name or a pen name; affiliations are optional and marked as self-reported. AI, Lean, endorsements and funding never decide whether you may publish.
The public service verifies that you can receive account email — any address works, not only academic ones. Private development instances may use a different account flow; the page shows which mode is running.
Public content and private account data
On the public service, text, versions, public author profiles, comments, collaboration records and every attachment uploaded to a work are open to readers, including original .lean files; whether a pinned version includes an attachment does not change whether it can be read on the work page. Only upload material you have the right to publish — never tokens, passwords, seed phrases, private emails or other people’s personal data without permission. When an author withdraws a work, a statement is recorded and earlier versions, pinned pinned versions and discussion remain.
Email addresses, password hashes, sessions and verification codes are stored on the server as private account data and never enter public profiles or works exports. Private conduct reports, Lean enrollments and pre-check reports have separate access controls. Backups may contain this private data and are kept as private operational data.
Attachments are stored as data; no code in them runs when you publish. Voluntary Lean checks process only authorized, fixed material in a separate restricted environment.
Resource limits during Alpha
During this prototype, each attachment is limited to 4 MiB, each version to 16 attachments, and the instance to 64 MiB of research attachments in total; works records also have an overall cap. When capacity is reached, new records or attachments may be paused and the page will show an error. Existing works and discussion remain readable.
Sign-in requests are rate-limited, and the public entry point also limits concurrency and request size. Ordinary academic debate, unconventional views, beginner mistakes and skepticism about AI are never grounds for conduct action. Any change to resource rules should state its scope and reason.
This is an early Alpha. Maintainers handle incidents and the voluntary check queue by hand, with no round-the-clock support or long-term availability guarantee. Keep your own copies of your text and materials, and use version citations and exports to preserve your records.
How to report problems
Report privacy, harassment, spam or malicious-code issues through the private conduct report on a work page, pointing to a specific version, comment or attachment; the author concerned can respond and request a limited appeal. Reports never automatically hide works, nor do they judge academic merit. Raise academic objections in the discussion, with your argument, scope and evidence.
For account or page failures, note the time, the page path and the error text (without any credentials). A complete appeals and account-action process with independent people is still being built; different operator IDs in the software are not presented as independent people.