- TeX 100%
| Filename | Latest commit message | Latest commit date |
|---|---|---|
|
|
||
| .github/workflows | ||
| corpus | ||
| docs | ||
| scripts | ||
| .gitattributes | ||
| .gitignore | ||
| app.js | ||
| index.html | ||
| Open TeX Reader.command | ||
| parser.js | ||
| parser.test.mjs | ||
| README.md | ||
| server.py | ||
| styles.css | ||
| test_server.py | ||
TeX Reader
Read the ideas without installing a TeX distribution.
TeX Reader is a small local browser app that turns .tex source into readable prose. It keeps headings, paragraphs, lists, quotations, code, and approximate equations while removing typesetting instructions. It includes a searchable collection of 4,887 TeX files from the active research projects and the organized Downloads collection.
The application uses Python's standard library and browser JavaScript. No MacTeX, npm packages, account, external fonts, or cloud service are required.
Start reading
git clone https://github.com/8b-is/tex-reader.git
cd tex-reader
python3 server.py --open
On macOS, you can also double-click Open TeX Reader.command. Keep its Terminal window open while reading; Control-C stops the server.
Requires Python 3.9 or newer and a modern browser. The app opens at http://127.0.0.1:8766/ by default. If that port is occupied, it chooses a free port and prints the address. It listens only on localhost.
Features
- Search the collection by filename, path, and project.
- Follow section headings through a document outline.
- Assemble available local chapter includes into one reading view.
- Switch between simplified prose and the original source.
- Change text size and light/dark appearance.
- Open an individual
.texfile from the browser. - Save the simplified document as a plain-text file.
The default starting paper is From Verification to Commitment. The reading guide provides an overview and suggested entry points across the collection.
Included collection
| Folder | TeX files | Contents |
|---|---|---|
corpus/research/ |
4,440 | Research sources, retaining their project-relative paths |
corpus/downloads/ |
447 | The earlier Downloads collection, retaining its original paths |
| Total | 4,887 | Drafts, fragments, alternate versions, and exact duplicates are preserved |
The collection excludes forks, .mothballed, Git internals, and dependency directories. These are source-file counts, not counts of unique completed papers. Images, PDFs, bibliographies, and TeX style/class files are not included; the reader does not require a PDF build.
corpus/manifest.json records every TeX path, byte size, and SHA-256. Available top-level license/notice files from the source projects are preserved beside the relevant corpus folders. See provenance and rights; this repository does not assign one new license to the collected papers.
Use your own folders
python3 server.py --open --root /path/to/papers --root /another/folder
Folders are indexed on startup. Restart the server to pick up newly added files. Individual browser-selected files can be read immediately, but their neighboring chapter files are not automatically available to the browser.
Reading limits
This is a best-effort source reader, not a TeX interpreter. Complex equations are approximations, tables are flattened, diagrams are omitted, and citations may use source keys. Arbitrary macros and conditional TeX programs are not fully evaluated. Reading notes identify several of these cases, and Source lets you inspect the original file.
Explicit input, include, subfile, and import references are assembled only within the same project. Missing files and include limits are reported. Source files are read-only: the app does not rewrite them, upload them, or execute TeX or shell commands embedded in them.
Development and verification
python3 -m unittest -v test_server.py
node --test parser.test.mjs
python3 scripts/verify_corpus.py
Node 22 or newer is used for parser tests; it is not needed to run the reader. Tests cover source parsing, real papers, chapter expansion, circular includes, literal examples, and project boundaries.
| File | Purpose |
|---|---|
server.py |
Local inventory, source reading, and chapter assembly |
parser.js |
TeX-to-readable-text parser |
index.html, app.js, styles.css |
Browser reading interface |
corpus/ |
Bundled TeX collection and integrity manifest |
docs/ |
Reading guide and provenance notes |