README Generator
Draft a repository README from project facts, then check setup and usage instructions plus licensing and contribution guidance before publishing.{{ summaryTitle }}
{{ summaryLine }}
{{ draftExportStatus }}
{{ computation.values.markdown }}
| Area | Status | Signal | Next step | Copy |
|---|---|---|---|---|
| {{ row.area }} | {{ row.status }} | {{ row.signal }} | {{ row.next_copy }} |
| Repository item | Status | Why it matters | Next action | Copy |
|---|---|---|---|---|
| {{ row.item }} | {{ row.status }} | {{ row.reason }} | {{ row.next_action }} |
A repository README is often the first explanation a user, contributor, or future maintainer sees. Its main job is to help that reader decide what the project is, whether it fits their need, and how to reach a first successful result. A project can have excellent code and still be hard to adopt when the setup assumptions or usage path are missing.
The most useful opening names the project, states the problem it solves, and gives enough context to identify the intended audience. Setup instructions then bridge the gap between interest and a working installation. A short, realistic usage example is stronger than a long feature list because it proves what success looks like.
Different repositories need different supporting sections. A command-line tool benefits from options and shell examples; a service needs endpoints and operations notes; a data project needs provenance and reproducibility guidance. Public projects also need clear contribution, conduct, support, security-reporting, and licensing routes. Internal projects may replace community guidance with ownership and operational expectations.
- Readers need orientation: purpose, audience, status, and the fastest useful example.
- Operators need repeatability: prerequisites, installation commands, configuration, testing, and support.
- Contributors need boundaries: contribution rules, conduct expectations, security reporting, and reuse terms.
A README should remain an entry point rather than absorbing every design note or tutorial. Longer material belongs in dedicated documentation, while the README keeps the quick path accurate and links to the deeper source.
Generated Markdown is a draft, not evidence that the repository actually contains the commands, workflow files, license text, or policy documents it names. Every link and instruction must be checked against the repository before publication.
How to Use This Tool:
Choose the profile that matches the repository's real job, then supply facts that a new reader can verify.
- Select the README profile and enter the exact Project name, a useful short description, and the repository URL when one exists.
- Add concrete features, prerequisites, installation commands, and one usage example that reaches a successful result. Use the profile-specific fields for API, options, data, or operations notes.
- Choose the real project status, license label, and contribution posture. Add support and security-policy paths only when those destinations will exist in the repository.
- Read the Section audit and Repository checklist, correct missing or weak facts, then inspect the generated Markdown before committing it as
README.md.
Interpreting Results:
The readiness percentage is a local coverage heuristic. A high score means the draft includes the weighted facts this generator checks; it does not prove that commands work, links resolve, badges report the right workflow, or legal terms match the project.
- Treat Missing as a gap in a high-value reader path such as installation or usage.
- Treat Check as a prompt for human judgment, especially for short descriptions, incomplete community files, non-GitHub repository URLs, and vague custom licenses.
- Use the repository checklist to create or verify companion files. The draft links to those files but does not create them.
Technical Details:
README generation is a structured Markdown transformation followed by a weighted content audit. Blank optional inputs suppress their sections, while required reader paths receive placeholder or missing-state guidance so the draft does not silently look complete.
Transformation Core:
The selected profile changes the profile-specific section title, not the underlying Markdown rules. Line-based facts become bullets, commands become fenced shell blocks, and environment-variable pairs become a table.
| Profile | Profile section | Emphasis |
|---|---|---|
| Library/package | API Reference | Public interfaces and integration notes |
| CLI tool | Options | Commands, flags, and shell behavior |
| Web app/API | API | Endpoints and service use |
| Data/research | Data and Reproducibility | Data sources, methods, and repeatable results |
| Internal service | Operations | Ownership, runtime, and support expectations |
Non-active status choices add a visible status note near the top. A table of contents appears only when requested and when the draft has at least five second-level sections. GitHub CI badges require a recognizable GitHub owner and repository path; other repository URLs do not receive that badge.
Rule Core:
| Audit area | Ready condition | Weight |
|---|---|---|
| Opening description | At least 60 characters | 16 |
| Installation | At least one command line | 16 |
| Usage | At least one example line | 16 |
| Project name and features | A name and at least three feature bullets | 20 |
| License, support, and community files | Specific reuse terms, a help route, and contribution files appropriate to the chosen posture | 21 |
| Requirements, profile notes, repository URL, security policy, and long-form docs | Evaluated as ready or optional according to the supplied facts | 29 |
Formula Core:
The readiness score is the rounded share of audit weight earned. Optional or partial rows can earn some credit, so it is not simply the percentage of rows marked Ready.
The displayed value is a whole-number percentage. The draft also counts second-level sections, whitespace-delimited words, lines, and up to the first four audit warnings, with an additional warning when badges are requested for a non-GitHub repository URL.
Accuracy and Privacy Notes:
Project facts are processed in the browser; the generator does not inspect the repository or send its contents to a server. That privacy boundary also explains the main accuracy limit: every generated claim comes from the text entered by the user.
- Run installation, test, and usage commands in a clean environment before publishing them.
- Confirm every relative path and external URL, especially contribution, conduct, security, documentation, and support links.
- A license label is not a license grant. Add the matching full license text and obtain appropriate legal review for custom or proprietary terms.
- Generated badges assume a GitHub Actions workflow named
ci.yml; change or remove the badge when the repository uses another workflow.
References:
- About the repository README file, GitHub Docs.
- Setting guidelines for repository contributors, GitHub Docs.
- CommonMark specification, CommonMark.
- SPDX License List, Software Package Data Exchange.