Linea documentation

Source for docs.linea.build, the official Linea documentation site. The site is built with Docusaurus v3 and deployed on Vercel.

• Live site: docs.linea.build
• Documentation for builders
• Lineth monorepo (open-source stack source)
• Linea and Lineth explained
• Community forum
• X
• Create an issue

Linea is a developer-ready layer 2 network, scaling Ethereum by providing an Ethereum-equivalent environment in which to execute transactions, which are then submitted to Ethereum Mainnet through a zero-knowledge rollup.

This documentation repository is built using Docusaurus, and the site itself is published at docs.linea.build.

This repository is the source for docs.linea.build, which documents both Linea and Lineth:

• Linea is the public, permissionless zkEVM Layer 2 network, along with its brand, token, tokenomics, and governance.
• Lineth (formerly the Linea Stack) is the open-source ZK-rollup stack and technical protocol that powers Linea Mainnet and other networks built on it. Lineth is an enterprise-grade EVM layer 2 framework for developing public, permissioned or private networks.

Linea Mainnet is built on Lineth. For the authoritative distinction, see Linea and Lineth. The Lineth source code lives in the LFDT-Lineth monorepo.

The site is organized into four documentation tabs plus a changelog, each mapping to a top-level directory under docs/:

The directories you are most likely to touch:

doc.linea/
├── docs/
│   ├── network/         # Linea Mainnet tab
│   ├── protocol/        # Protocol tab: the Lineth protocol (includes linea-vs-lineth.mdx and the glossary)
│   ├── stack/           # Lineth Stack tab: the open-source stack operators deploy
│   ├── api/             # Reference tab: Linea (linea-smart-contracts/ is auto-generated, don't edit)
│   └── changelog/       # Releases across Lineth and the Linea network, plus the security council record
├── static/img/          # Images and SVGs, mirroring the docs/ structure
└── src/                 # Site components, theme, and CSS (see "Working with AI agents")

Build configuration, the scripts/ codegen helpers (social cards, categories, the llms.txt pipeline), the api/ Vercel serverless functions (the docs feedback endpoint), sidebars.js, redirects.json, vercel.json, and the agent tooling (AGENTS.md, .cursor/rules/, .agents/skills/) all live at the repo root.

• Node.js 24.18.0 (pinned in .nvmrc and engines in package.json).
• npm as the package manager.

It is highly recommended that you use a tool like nvm to manage Node.js versions. From the root of the project, run nvm install followed by nvm use to install and select the version specified in .nvmrc.

You need Node.js installed to run live previews of the docs locally. See Requirements above.

1. Install dependencies (only needed once, or when dependencies change):

   npm install

2. Start the dev server with hot reload:

   npm run start

3. To produce and preview a production build:

   npm run build
   npm run serve

   npm run build generates static content into the build directory, which can be served by any static host.

If you hit stale cache or build issues, npm run clear clears the Docusaurus cache, and npm run clean-slate performs a fuller reset.

This documentation applies the Consensys style guide, other than for images and other media, for which we have specific guidelines.

Detailed, machine-readable writing and formatting conventions (voice, terminology, Markdown/MDX syntax, page structure) live in AGENTS.md and .cursor/rules/. They apply to human and AI contributors alike, and Cursor auto-applies them when you edit a matching file.

Any images for articles should be added to the /static/img/ directory. Here, the /img folder structure mirrors that of the docs, and images are stored in the corresponding subdirectory. If you have added a new article in /docs, create the new subdirectory in /img according to this structure.

Use the following JSX formatting to add an image into an article:

<div class="center-container">
  <div class="img-medium">
      <img
        src="/img/path-to-file/image.png"
        alt="alt text here"
      />
  </div>
</div>

To adjust the display size of the image, change the size class to one of: img-xsmall, img-small, img-medium, img-large. Adhere to these predefined sizings where possible.

For SVG files, place the import statement at the top of the file, beneath the front matter:

import MySVGGraphic from "/img/path-to-file/image.svg";

<div className="img-large">
  <MySVGGraphic />
</div>

As above, you can adjust the display size of the SVG by changing the class applied.

A social card for any new article will be generated and applied when the site builds, so there is no need to manually create or add one.

See CONTRIBUTING.md for the full contribution workflow, including how to contribute to the glossary and additional Consensys docs resources.

Run these locally before opening or updating a PR, to mirror what CI runs:

npm run lint        # ESLint + Stylelint
npm run lint:style  # CSS only
npm run typecheck   # TypeScript

Spelling and style (Vale) and link checking run automatically in CI on your PR. The main linter, Vale, checks spelling, formatting, and adherence to various style guides. Please pay attention to its suggestions and warnings: these will not prevent your PR from building, but you should address them nevertheless. See our guidance on using Vale for more information.

The following GitHub Actions workflows run on pull requests:

This repo follows Conventional Commits. Use a standard type and a short, imperative description, for example:

docs: clarify bridging guide

The repo configures commitlint (@commitlint/config-conventional, standard types) and uses lint-staged with Prettier for formatting. Run npm run lint and npm run format locally before pushing so your changes are clean. Published release notes live as documentation content under docs/changelog.

The site is deployed on Vercel:

• Every pull request gets a unique Vercel preview URL so you can review your changes in context.
• Production deploys automatically from the main branch.

AGENTS.md at the repo root is the canonical guidance for AI coding agents and human contributors alike: critical rules, plus an index of the detailed rule files in .cursor/rules/ (editorial voice, terminology, content structure, Markdown formatting, contributor workflow, CSS/design tokens) and the packaged skills in .agents/skills/ (authoring a new page, pre-PR content review, dependency maintenance).

Separately from the rules above, the site publishes a machine-readable copy of itself for LLM consumers. On build, the postbuild step (scripts/agent-docs/generate.js) emits an llms.txt index and a Markdown version of each route. Vercel serves these via content negotiation: a request with an Accept: text/markdown header receives the Markdown variant of a page rather than HTML. The llms.txt file and a /.well-known/api-catalog are advertised through response Link headers (see vercel.json).

Validate this pipeline locally with npm run agent-docs:check and npm run agent-docs:verify-preview.

See something missing? Error in our documentation? Create an issue.

Alternatively, help us improve our documentation. See CONTRIBUTING.md for the full contribution workflow (including guidance for ecosystem contributions), or get in touch via the community forum or Linea Support.

This project is licensed under the terms of the LICENSE file.