Skip to main content

MachineChecked

Read a mathematical paper beside its Lean formalisation.

Select a statement to inspect its Lean counterpart, proof results and correspondence reviews. MachineChecked runs locally in your Lean project and saves the links you choose.

Try the demo · Browse papers · Setup guide

Open your paper

From your Lean project's directory:

lake build
uvx machinechecked view

MachineChecked finds the paper and opens the reader. You can also name a LaTeX file, PDF or URL:

uvx machinechecked view paper/main.tex

Requires Python 3.11+, your project's Lean toolchain, and TeX Live for LaTeX or Poppler for PDFs. The web reader is included. For a persistent command, run uv tool install machinechecked; then use machinechecked directly.

Select a claim from the contents or paper margin. Expand comparison shows the paper, complete readable statement and exact Lean together. Follow existing \lean{...} tags, confirm a suggestion, or choose a declaration. The reader saves your choices in machinechecked.links.json. Projects with a formalization.yaml supply their paper sources and declared links automatically.

Read the evidence

A Lean proof establishes a formal proposition. A correspondence review compares that proposition and its definitions with the paper. Source checks identify the audited text; independent kernel replay checks the proof again. The reader shows each result and its provenance in Audit details.

The optional Review view guides you through three steps:

  1. Read the paper and record your interpretation.
  2. Compare the Lean statement's assumptions and conclusion; record a verdict.
  3. Read the author's review, save your report in the browser, and export it.

The maintainer evaluates reports and promotes reviews into the project's acceptance records. Review guide →

Share the reader

uvx machinechecked build                 # .machinechecked/MachineChecked.html
uvx machinechecked build --site public/  # a static site with audit.json
uvx machinechecked check                 # check linked declarations for automation

The HTML file opens directly in a browser. It includes the paper, Lean statements and audit data; review those contents before publishing. Host the site directory on GitHub Pages or another static host. Published audits carry their author's name, project commit and paper revision.

Copy link shares a claim. Report issue opens a draft in the audit's GitHub repository, with the claim identifier, declarations and commits filled in. Set that repository with --issues owner/repo.

With a coding agent

machinechecked check --json prints structured results for scripts and agents. The MachineChecked skill covers commands, status codes and evidence. Agent guide →

Full assessment

Projects with locked specifications, a claim inventory and review records can configure machinechecked.toml. Full assessment adds exact-type checks, definition-dependent reviews, witnesses, refutations and independent kernel replay. Configuration guide →

Documentation

Apache-2.0 licensed.

Metadata

Release files for machinechecked 0.1.15

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for machinechecked 0.1.15
File Size Uploaded Provenance
machinechecked-0.1.15.tar.gz 1.1 MB With provenance Details

Built distribution (wheel)

Table of built distributions (wheels) for machinechecked 0.1.15
File Interpreter ABI Platform
machinechecked-0.1.15-py3-none-any.whl Python 3 none any Details

Total release size: 2.2 MB

Release files / machinechecked-0.1.15.tar.gz

Download URL machinechecked-0.1.15.tar.gz
Size 1.1 MB
Tags Source
SHA-256 checksum
How to use checksums
56645d2db32b6a6b152126434f12b788304df2e3543fe4c1d21aeaef31a8abdb
BLAKE2b-256 checksum
How to use checksums
5682cd5be92c45eadccefed9e84133a301f4bb3f5200cb31a78a155cd32ad58a
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 19, 2026.

Transparency log

Release files / machinechecked-0.1.15-py3-none-any.whl

Download URL machinechecked-0.1.15-py3-none-any.whl
Size 1.1 MB
Tags Python 3
SHA-256 checksum
How to use checksums
d5d4e937265fbd098b9c77613dea757ef9ed784e22275594e4dfa6240e61d379
BLAKE2b-256 checksum
How to use checksums
a34c55d7afff785ee0b53e932450d39ceec33690e74ef4dc9cc538830e8ef530
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 19, 2026.

Transparency log

Security for machinechecked 0.1.15

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

Full provenance

All 2 files in this release have been uploaded with provenance metadata.

Select from 2 files

Attestation for machinechecked-0.1.15.tar.gz

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 19, 2026.

Transparency log

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page