GETTING STARTED FUNDAMENTALS

Configure Hugo

Configuration file

Create a site configuration file in the root of your project directory, naming it hugo.toml, hugo.yaml, or hugo.json, with that order of precedence.

my-project/
└── hugo.toml

A simple example:

hugo.
      
baseURL: https://example.org/
languageCode: en-us
params:
  contact:
    email: info@example.org
    phone: +1 202-555-1212
  subtitle: The Best Widgets on Earth
title: ABC Widgets, Inc.

baseURL = 'https://example.org/'
languageCode = 'en-us'
title = 'ABC Widgets, Inc.'
[params]
  subtitle = 'The Best Widgets on Earth'
  [params.contact]
    email = 'info@example.org'
    phone = '+1 202-555-1212'

{
   "baseURL": "https://example.org/",
   "languageCode": "en-us",
   "params": {
      "contact": {
         "email": "info@example.org",
         "phone": "+1 202-555-1212"
      },
      "subtitle": "The Best Widgets on Earth"
   },
   "title": "ABC Widgets, Inc."
}

To use a different configuration file when building your site, use the --config flag:

hugo --config other.toml

Combine two or more configuration files, with left-to-right precedence:

hugo --config a.toml,b.yaml,c.json

Configuration directory

Instead of a single site configuration file, split your configuration by environment, root configuration key, and language. For example:

my-project/
└── config/
    ├── _default/
    │   ├── hugo.toml
    │   ├── menus.en.toml
    │   ├── menus.de.toml
    │   └── params.toml
    └── production/
        └── params.toml

The root configuration keys are build, caches, cascade, deployment, frontmatter, imaging, languages, markup, mediatypes, menus, minify, module, outputformats, outputs, params, permalinks, privacy, related, security, segments, server, services, sitemap, and taxonomies.

Omit the root key

When splitting the configuration by root key, omit the root key in the given file. For example, these are equivalent:

hugo.
      
params:
  foo: bar

[params]
  foo = 'bar'

{
   "params": {
      "foo": "bar"
   }
}
params.
      
foo: bar

foo = 'bar'

{
   "foo": "bar"
}

Recursive parsing

Hugo parses the config directory recursively, allowing you to organize the files into subdirectories. For example:

my-project/
└── config/
    └── _default/
        ├── navigation/
        │   ├── menus.de.toml
        │   └── menus.en.toml
        └── hugo.toml

Example 

my-project/
└── config/
    ├── _default/
    │   ├── hugo.toml
    │   ├── menus.en.toml
    │   ├── menus.de.toml
    │   └── params.toml
    ├── production/
    │   ├── hugo.toml
    │   └── params.toml
    └── staging/
        ├── hugo.toml
        └── params.toml

Considering the structure above, when running hugo --environment staging, Hugo will use every setting from config/_default and merge staging’s on top of those.

Let’s take an example to understand this better. Let’s say you are using Google Analytics for your website. This requires you to specify a Google tag ID in your site configuration:

hugo.
      
services:
  googleAnalytics:
    ID: G-XXXXXXXXX

[services]
  [services.googleAnalytics]
    ID = 'G-XXXXXXXXX'

{
   "services": {
      "googleAnalytics": {
         "ID": "G-XXXXXXXXX"
      }
   }
}

Now consider the following scenario:

  1. You don’t want to load the analytics code when running hugo server.

  2. You want to use different Google tag IDs for your production and staging environments. For example:

    • G-PPPPPPPPP for production
    • G-SSSSSSSSS for staging

To satisfy these requirements, configure your site as follows:

  1. config/_default/hugo.toml

    Exclude the services.googleAnalytics section. This will prevent loading of the analytics code when you run hugo server.

    By default, Hugo sets its environment to development when running hugo server. In the absence of a config/development directory, Hugo uses the config/_default directory.

  2. config/production/hugo.toml

    Include this section only:

    hugo.
          
    services:
  googleAnalytics:
    ID: G-PPPPPPPPP

    [services]
  [services.googleAnalytics]
    ID = 'G-PPPPPPPPP'

    {
   "services": {
      "googleAnalytics": {
         "ID": "G-PPPPPPPPP"
      }
   }
}

    You do not need to include other parameters in this file. Include only those parameters that are specific to your production environment. Hugo will merge these parameters with the default configuration.

    By default, Hugo sets its environment to production when running hugo. The analytics code will use the G-PPPPPPPPP tag ID.

  3. config/staging/hugo.toml

    Include this section only:

    hugo.
          
    services:
  googleAnalytics:
    ID: G-SSSSSSSSS

    [services]
  [services.googleAnalytics]
    ID = 'G-SSSSSSSSS'

    {
   "services": {
      "googleAnalytics": {
         "ID": "G-SSSSSSSSS"
      }
   }
}

    You do not need to include other parameters in this file. Include only those parameters that are specific to your staging environment. Hugo will merge these parameters with the default configuration.

    To build your staging site, run hugo --environment staging. The analytics code will use the G-SSSSSSSSS tag ID.

Merge configuration from themes

The configuration value for _merge can be one of:

none
No merge.
shallow
Only add values for new keys.
deep
Add values for new keys, merge existing.

Note that you don’t need to be so verbose as in the default setup below; a _merge value higher up will be inherited if not set.

hugo.
      
build:
  _merge: none
caches:
  _merge: none
cascade:
  _merge: none
deployment:
  _merge: none
frontmatter:
  _merge: none
httpcache:
  _merge: none
imaging:
  _merge: none
languages:
  _merge: none
  en:
    _merge: none
    menus:
      _merge: shallow
    params:
      _merge: deep
markup:
  _merge: none
mediatypes:
  _merge: shallow
menus:
  _merge: shallow
minify:
  _merge: none
module:
  _merge: none
outputformats:
  _merge: shallow
outputs:
  _merge: none
pagination:
  _merge: none
params:
  _merge: deep
permalinks:
  _merge: none
privacy:
  _merge: none
related:
  _merge: none
security:
  _merge: none
segments:
  _merge: none
server:
  _merge: none
services:
  _merge: none
sitemap:
  _merge: none
taxonomies:
  _merge: none

[build]
  _merge = 'none'
[caches]
  _merge = 'none'
[cascade]
  _merge = 'none'
[deployment]
  _merge = 'none'
[frontmatter]
  _merge = 'none'
[httpcache]
  _merge = 'none'
[imaging]
  _merge = 'none'
[languages]
  _merge = 'none'
  [languages.en]
    _merge = 'none'
    [languages.en.menus]
      _merge = 'shallow'
    [languages.en.params]
      _merge = 'deep'
[markup]
  _merge = 'none'
[mediatypes]
  _merge = 'shallow'
[menus]
  _merge = 'shallow'
[minify]
  _merge = 'none'
[module]
  _merge = 'none'
[outputformats]
  _merge = 'shallow'
[outputs]
  _merge = 'none'
[pagination]
  _merge = 'none'
[params]
  _merge = 'deep'
[permalinks]
  _merge = 'none'
[privacy]
  _merge = 'none'
[related]
  _merge = 'none'
[security]
  _merge = 'none'
[segments]
  _merge = 'none'
[server]
  _merge = 'none'
[services]
  _merge = 'none'
[sitemap]
  _merge = 'none'
[taxonomies]
  _merge = 'none'

{
   "build": {
      "_merge": "none"
   },
   "caches": {
      "_merge": "none"
   },
   "cascade": {
      "_merge": "none"
   },
   "deployment": {
      "_merge": "none"
   },
   "frontmatter": {
      "_merge": "none"
   },
   "httpcache": {
      "_merge": "none"
   },
   "imaging": {
      "_merge": "none"
   },
   "languages": {
      "_merge": "none",
      "en": {
         "_merge": "none",
         "menus": {
            "_merge": "shallow"
         },
         "params": {
            "_merge": "deep"
         }
      }
   },
   "markup": {
      "_merge": "none"
   },
   "mediatypes": {
      "_merge": "shallow"
   },
   "menus": {
      "_merge": "shallow"
   },
   "minify": {
      "_merge": "none"
   },
   "module": {
      "_merge": "none"
   },
   "outputformats": {
      "_merge": "shallow"
   },
   "outputs": {
      "_merge": "none"
   },
   "pagination": {
      "_merge": "none"
   },
   "params": {
      "_merge": "deep"
   },
   "permalinks": {
      "_merge": "none"
   },
   "privacy": {
      "_merge": "none"
   },
   "related": {
      "_merge": "none"
   },
   "security": {
      "_merge": "none"
   },
   "segments": {
      "_merge": "none"
   },
   "server": {
      "_merge": "none"
   },
   "services": {
      "_merge": "none"
   },
   "sitemap": {
      "_merge": "none"
   },
   "taxonomies": {
      "_merge": "none"
   }
}

All configuration settings

archetypeDir

(string) The directory where Hugo finds archetype files (content templates). Default is archetypes. Also see Module Mounts Config for an alternative way to configure this directory (from Hugo 0.56).

assetDir

(string) The directory where Hugo finds asset files used in Hugo Pipes. Default is assets. Also see Module Mounts Config for an alternative way to configure this directory (from Hugo 0.56).

baseURL

(string) The absolute URL (protocol, host, path, and trailing slash) of your published site (e.g., https://www.example.org/docs/).

build

See Configure Build.

buildDrafts

(bool) Include drafts when building. Default is false.

buildExpired

(bool) Include content already expired. Default is false.

buildFuture

(bool) Include content with a future publication date. Default is false.

caches

See Configure File Caches.

capitalizeListTitles
New in v0.123.3

(bool) Whether to capitalize automatic list titles. Applicable to section, taxonomy, and term pages. Default is true. You can change the capitalization style in your site configuration to one of ap, chicago, go, firstupper, or none. See details.

cascade

Pass down default configuration values (front matter) to pages in the content tree. The options in site config is the same as in page front matter, see Front Matter Cascade.

canonifyURLs

(bool) See details before enabling this feature. Default is false.

cleanDestinationDir

(bool) When building, removes files from destination not found in static directories. Default is false.

contentDir

(string) The directory from where Hugo reads content files. Default is content. Also see Module Mounts Config for an alternative way to configure this directory (from Hugo 0.56).

(string) Copyright notice for your site, typically displayed in the footer.

dataDir

(string) The directory from where Hugo reads data files. Default is data. Also see Module Mounts Config for an alternative way to configure this directory (from Hugo 0.56).

defaultContentLanguage

(string) Content without language indicator will default to this language. Default is en.

defaultContentLanguageInSubdir

(bool) Render the default content language in subdir, e.g. content/en/. The site root / will then redirect to /en/. Default is false.

disableAliases

(bool) Will disable generation of alias redirects. Note that even if disableAliases is set, the aliases themselves are preserved on the page. The motivation with this is to be able to generate 301 redirects in an .htaccess, a Netlify _redirects file or similar using a custom output format. Default is false.

disableHugoGeneratorInject

(bool) Hugo will, by default, inject a generator meta tag in the HTML head on the home page only. You can turn it off, but we would really appreciate if you don’t, as this is a good way to watch Hugo’s popularity on the rise. Default is false.

disableKinds

(string slice) Disable rendering of the specified page kinds, any of 404, home, page, robotstxt, rss, section, sitemap, taxonomy, or term.

disableLiveReload

(bool) Disable automatic live reloading of browser window. Default is false.

disablePathToLower

(bool) Do not convert the url/path to lowercase. Default is false.

enableEmoji

(bool) Enable Emoji emoticons support for page content; see the emoji shortcode quick reference guide. Default is false.

enableGitInfo

(bool) Enable .GitInfo object for each page (if the Hugo site is versioned by Git). This will then update the Lastmod parameter for each page using the last git commit date for that content file. Default is false.

enableMissingTranslationPlaceholders

(bool) Show a placeholder instead of the default value or an empty string if a translation is missing. Default is false.

enableRobotsTXT

(bool) Enable generation of robots.txt file. Default is false.

frontmatter

See Front matter Configuration.

hasCJKLanguage

(bool) If true, auto-detect Chinese/Japanese/Korean Languages in the content. This will make .Summary and .WordCount behave correctly for CJK languages. Default is false.

imaging

See image processing configuration.

languageCode

(string) A language tag as defined by RFC 5646. This value is used to populate:

When present in the root of the configuration, this value is ignored if one or more language keys exists. Please specify this value independently for each language key.

languages

See Configure Languages.

disableLanguages

See Disable a Language

markup

See Configure Markup.

mediaTypes

See Configure Media Types.

See Menus.

minify

See Configure Minify.

module

Module configuration see module configuration.

newContentEditor

(string) The editor to use when creating new content.

noChmod

(bool) Don’t sync permission mode of files. Default is false.

noTimes

(bool) Don’t sync modification time of files. Default is false.

outputFormats

See custom output formats.

pagination

See configure pagination.

See Content Management.

pluralizeListTitles

(bool) Whether to pluralize automatic list titles. Applicable to section pages. Default is true.

publishDir

(string) The directory to where Hugo will write the final static site (the HTML files etc.). Default is public.

refLinksErrorLevel

(string) When using ref or relref to resolve page links and a link cannot be resolved, it will be logged with this log level. Valid values are ERROR (default) or WARNING. Any ERROR will fail the build (exit -1). Default is ERROR.

refLinksNotFoundURL

(string) URL to be used as a placeholder when a page reference cannot be found in ref or relref. Is used as-is.

See Related Content.

relativeURLs

(bool) See details before enabling this feature. Default is false.

renderSegments
New in v0.124.0

(string slice) A list of segments to render. If not set, everything will be rendered. This is more commonly set in a CLI flag, e.g. hugo --renderSegments segment1,segment2. The segment names must match the names in the segments configuration.

removePathAccents

(bool) Removes non-spacing marks from composite characters in content paths. Default is false.

content/post/hügó.md → https://example.org/post/hugo/
sectionPagesMenu

See Menus.

security

See Security Policy.

segments

See Segments.

sitemap

Default sitemap configuration.

summaryLength

(int) Applicable to automatic summaries, the approximate number of words to render when calling the Summary method on a Page object. Default is 70.

taxonomies

See Configure Taxonomies.

theme

See module configuration for how to import a theme.

themesDir

(string) The directory where Hugo reads the themes from. Default is themes.

timeout

(string) Timeout for generating page contents, specified as a duration or in seconds. Note: this is used to bail out of recursive content generation. You might need to raise this limit if your pages are slow to generate (e.g., because they require large image processing or depend on remote contents). Default is 30s.

timeZone

(string) The time zone (or location), e.g. Europe/Oslo, used to parse front matter dates without such information and in the time function. The list of valid values may be system dependent, but should include UTC, Local, and any location in the IANA Time Zone database.

title

(string) Site title.

titleCaseStyle

(string) Default is ap. See Configure Title Case.

uglyURLs

(bool) When enabled, creates URL of the form /filename.html instead of /filename/. Default is false.

watch

(bool) Watch filesystem for changes and recreate as needed. Default is false.

Configure build

The build configuration section contains global build-related configuration options.

hugo.
      
build:
  buildStats:
    disableClasses: false
    disableIDs: false
    disableTags: false
    enable: false
  cacheBusters:
  - source: (postcss|tailwind)\.config\.js
    target: (css|styles|scss|sass)
  noJSConfigInAssets: false
  useResourceCacheWhen: fallback

[build]
  noJSConfigInAssets = false
  useResourceCacheWhen = 'fallback'
  [build.buildStats]
    disableClasses = false
    disableIDs = false
    disableTags = false
    enable = false
  [[build.cacheBusters]]
    source = '(postcss|tailwind)\.config\.js'
    target = '(css|styles|scss|sass)'

{
   "build": {
      "buildStats": {
         "disableClasses": false,
         "disableIDs": false,
         "disableTags": false,
         "enable": false
      },
      "cacheBusters": [
         {
            "source": "(postcss|tailwind)\\.config\\.js",
            "target": "(css|styles|scss|sass)"
         }
      ],
      "noJSConfigInAssets": false,
      "useResourceCacheWhen": "fallback"
   }
}
buildStats New in v0.115.1
When enabled, creates a hugo_stats.json file in the root of your project. This file contains arrays of the class attributes, id attributes, and tags of every HTML element within your published site. Use this file as data source when removing unused CSS from your site. This process is also known as pruning, purging, or tree shaking.

Exclude class attributes, id attributes, or tags from hugo_stats.json with the disableClasses, disableIDs, and disableTags keys.

Due to the nature of partial server builds, new HTML entities are added while the server is running, but old values will not be removed until you restart the server or run a regular hugo build.

cachebusters
See Configure Cache Busters
noJSConfigInAssets
Turn off writing a jsconfig.json into your /assets folder with mapping of imports from running js.Build. This file is intended to help with intellisense/navigation inside code editors such as VS Code. Note that if you do not use js.Build, no file will be written.
useResourceCacheWhen
When to use the cached resources in /resources/_gen for PostCSS and ToCSS. Valid values are never, always and fallback. The last value means that the cache will be tried if PostCSS/extended version is not available.

Configure cache busters

New in v0.112.0

The build.cachebusters configuration option was added to support development using Tailwind 3.x’s JIT compiler where a build configuration may look like this:

hugo.
      
build:
  buildStats:
    enable: true
  cachebusters:
  - source: assets/watching/hugo_stats\.json
    target: styles\.css
  - source: (postcss|tailwind)\.config\.js
    target: css
  - source: assets/.*\.(js|ts|jsx|tsx)
    target: js
  - source: assets/.*\.(.*)$
    target: $1

[build]
  [build.buildStats]
    enable = true
  [[build.cachebusters]]
    source = 'assets/watching/hugo_stats\.json'
    target = 'styles\.css'
  [[build.cachebusters]]
    source = '(postcss|tailwind)\.config\.js'
    target = 'css'
  [[build.cachebusters]]
    source = 'assets/.*\.(js|ts|jsx|tsx)'
    target = 'js'
  [[build.cachebusters]]
    source = 'assets/.*\.(.*)$'
    target = '$1'

{
   "build": {
      "buildStats": {
         "enable": true
      },
      "cachebusters": [
         {
            "source": "assets/watching/hugo_stats\\.json",
            "target": "styles\\.css"
         },
         {
            "source": "(postcss|tailwind)\\.config\\.js",
            "target": "css"
         },
         {
            "source": "assets/.*\\.(js|ts|jsx|tsx)",
            "target": "js"
         },
         {
            "source": "assets/.*\\.(.*)$",
            "target": "$1"
         }
      ]
   }
}

When buildStats New in v0.115.1 is enabled, Hugo writes a hugo_stats.json file on each build with HTML classes etc. that’s used in the rendered output. Changes to this file will trigger a rebuild of the styles.css file. You also need to add hugo_stats.json to Hugo’s server watcher. See Hugo Starter Tailwind Basic for a running example.

source
A regexp matching file(s) relative to one of the virtual component directories in Hugo, typically assets/....
target
A regexp matching the keys in the resource cache that should be expired when source changes. You can use the matching regexp groups from source in the expression, e.g. $1.

Configure server

This is only relevant when running hugo server, and it allows to set HTTP headers during development, which allows you to test out your Content Security Policy and similar. The configuration format matches Netlify’s with slightly more powerful Glob matching:

hugo.
      
server:
  headers:
  - for: /**
    values:
      Content-Security-Policy: script-src localhost:1313
      Referrer-Policy: strict-origin-when-cross-origin
      X-Content-Type-Options: nosniff
      X-Frame-Options: DENY
      X-XSS-Protection: 1; mode=block

[server]
  [[server.headers]]
    for = '/**'
    [server.headers.values]
      Content-Security-Policy = 'script-src localhost:1313'
      Referrer-Policy = 'strict-origin-when-cross-origin'
      X-Content-Type-Options = 'nosniff'
      X-Frame-Options = 'DENY'
      X-XSS-Protection = '1; mode=block'

{
   "server": {
      "headers": [
         {
            "for": "/**",
            "values": {
               "Content-Security-Policy": "script-src localhost:1313",
               "Referrer-Policy": "strict-origin-when-cross-origin",
               "X-Content-Type-Options": "nosniff",
               "X-Frame-Options": "DENY",
               "X-XSS-Protection": "1; mode=block"
            }
         }
      ]
   }
}

Since this is “development only”, it may make sense to put it below the development environment:

config/development/server.
      
headers:
- for: /**
  values:
    Content-Security-Policy: script-src localhost:1313
    Referrer-Policy: strict-origin-when-cross-origin
    X-Content-Type-Options: nosniff
    X-Frame-Options: DENY
    X-XSS-Protection: 1; mode=block

[[headers]]
  for = '/**'
  [headers.values]
    Content-Security-Policy = 'script-src localhost:1313'
    Referrer-Policy = 'strict-origin-when-cross-origin'
    X-Content-Type-Options = 'nosniff'
    X-Frame-Options = 'DENY'
    X-XSS-Protection = '1; mode=block'

{
   "headers": [
      {
         "for": "/**",
         "values": {
            "Content-Security-Policy": "script-src localhost:1313",
            "Referrer-Policy": "strict-origin-when-cross-origin",
            "X-Content-Type-Options": "nosniff",
            "X-Frame-Options": "DENY",
            "X-XSS-Protection": "1; mode=block"
         }
      }
   ]
}

You can also specify simple redirects rules for the server. The syntax is again similar to Netlify’s.

Note that a status code of 200 will trigger a URL rewrite, which is what you want in SPA situations, e.g:

config/development/server.
      
redirects:
- force: false
  from: /myspa/**
  status: 200
  to: /myspa/

[[redirects]]
  force = false
  from = '/myspa/**'
  status = 200
  to = '/myspa/'

{
   "redirects": [
      {
         "force": false,
         "from": "/myspa/**",
         "status": 200,
         "to": "/myspa/"
      }
   ]
}

Setting force=true will make a redirect even if there is existing content in the path. Note that before Hugo 0.76 force was the default behavior, but this is inline with how Netlify does it.

404 server error page

New in v0.103.0

Hugo will, by default, render all 404 errors when running hugo server with the 404.html template. Note that if you have already added one or more redirects to your server configuration, you need to add the 404 redirect explicitly, e.g:

config/development/server.
      
redirects:
- from: /**
  status: 404
  to: /404.html

[[redirects]]
  from = '/**'
  status = 404
  to = '/404.html'

{
   "redirects": [
      {
         "from": "/**",
         "status": 404,
         "to": "/404.html"
      }
   ]
}

With a multilingual site, define the redirect for the default content language last:

config/development/server.
      
defaultContentLanguage: en
defaultContentLanguageInSubdir: false
redirects:
- from: /fr/**
  status: 404
  to: /fr/404.html
- from: /**
  status: 404
  to: /404.html

defaultContentLanguage = 'en'
defaultContentLanguageInSubdir = false
[[redirects]]
  from = '/fr/**'
  status = 404
  to = '/fr/404.html'
[[redirects]]
  from = '/**'
  status = 404
  to = '/404.html'

{
   "defaultContentLanguage": "en",
   "defaultContentLanguageInSubdir": false,
   "redirects": [
      {
         "from": "/fr/**",
         "status": 404,
         "to": "/fr/404.html"
      },
      {
         "from": "/**",
         "status": 404,
         "to": "/404.html"
      }
   ]
}

If you are serving the default content language from a subdirectory:

config/development/server.
      
defaultContentLanguage: en
defaultContentLanguageInSubdir: true
redirects:
- from: /fr/**
  status: 404
  to: /fr/404.html
- from: /**
  status: 404
  to: /en/404.html

defaultContentLanguage = 'en'
defaultContentLanguageInSubdir = true
[[redirects]]
  from = '/fr/**'
  status = 404
  to = '/fr/404.html'
[[redirects]]
  from = '/**'
  status = 404
  to = '/en/404.html'

{
   "defaultContentLanguage": "en",
   "defaultContentLanguageInSubdir": true,
   "redirects": [
      {
         "from": "/fr/**",
         "status": 404,
         "to": "/fr/404.html"
      },
      {
         "from": "/**",
         "status": 404,
         "to": "/en/404.html"
      }
   ]
}

Configure title case

By default, Hugo follows the capitalization rules published in the Associated Press Stylebook when creating automatic section titles, and when transforming strings with the strings.Title function.

Change this behavior by setting titleCaseStyle in your site configuration to any of the values below:

ap
Use the capitalization rules published in the Associated Press Stylebook.
chicago
Use the capitalization rules published in the Chicago Manual of Style.
go
Capitalize the first letter of every word.
firstupper
Capitalize the first letter of the first word.
none
Disable transformation of automatic section titles, and disable the transformation performed by the strings.Title function. This is useful if you would prefer to manually capitalize section titles as needed, and to bypass opinionated theme usage of the strings.Title function.

Configuration environment variables

DART_SASS_BINARY
(string) The absolute path to the Dart Sass executable. By default, Hugo searches for the executable in each of the paths in the PATH environment variable.
HUGO_ENVIRONMENT
(string) Overrides the default environment, typically one of development, staging, or production.
HUGO_FILE_LOG_FORMAT
(string) A format string for the file path, line number, and column number displayed when reporting errors, or when calling the Position method from a shortcode or Markdown render hook. Valid tokens are :file, :line, and :col. Default is :file::line::col.
New in v0.123.0
HUGO_MEMORYLIMIT
(int) The maximum amount of system memory, in gigabytes, that Hugo can use while rendering your site. Default is 25% of total system memory.
HUGO_NUMWORKERMULTIPLIER
(int) The number of workers used in parallel processing. Default is the number of logical CPUs.

Configure with environment variables

In addition to the 3 configuration options already mentioned, configuration key-values can be defined through operating system environment variables.

For example, the following command will effectively set a website’s title on Unix-like systems:

$ env HUGO_TITLE="Some Title" hugo

This is really useful if you use a service such as Netlify to deploy your site. Look at the Hugo docs Netlify configuration file for an example.

If you are using snake_cased variable names, the above will not work. Hugo determines the delimiter to use by the first character after HUGO. This allows you to define environment variables on the form HUGOxPARAMSxAPI_KEY=abcdefgh, using any allowed delimiter.

Ignore content and data files when rendering

To exclude specific files from the content, data, and i18n directories when rendering your site, set ignoreFiles to one or more regular expressions to match against the absolute file path.

To ignore files ending with .foo or .boo:

hugo.
      
ignoreFiles:
- \.foo$
- \.boo$

ignoreFiles = ['\.foo$', '\.boo$']

{
   "ignoreFiles": [
      "\\.foo$",
      "\\.boo$"
   ]
}

To ignore a file using the absolute file path:

hugo.
      
ignoreFiles:
- ^/home/user/project/content/test\.md$

ignoreFiles = ['^/home/user/project/content/test\.md$']

{
   "ignoreFiles": [
      "^/home/user/project/content/test\\.md$"
   ]
}

Configure front matter

Configure dates

Dates are important in Hugo, and you can configure how Hugo assigns dates to your content pages. You do this by adding a frontmatter section to your hugo.toml.

The default configuration is:

hugo.
      
frontmatter:
  date:
  - date
  - publishdate
  - pubdate
  - published
  - lastmod
  - modified
  expiryDate:
  - expirydate
  - unpublishdate
  lastmod:
  - :git
  - lastmod
  - modified
  - date
  - publishdate
  - pubdate
  - published
  publishDate:
  - publishdate
  - pubdate
  - published
  - date

[frontmatter]
  date = ['date', 'publishdate', 'pubdate', 'published', 'lastmod', 'modified']
  expiryDate = ['expirydate', 'unpublishdate']
  lastmod = [':git', 'lastmod', 'modified', 'date', 'publishdate', 'pubdate', 'published']
  publishDate = ['publishdate', 'pubdate', 'published', 'date']

{
   "frontmatter": {
      "date": [
         "date",
         "publishdate",
         "pubdate",
         "published",
         "lastmod",
         "modified"
      ],
      "expiryDate": [
         "expirydate",
         "unpublishdate"
      ],
      "lastmod": [
         ":git",
         "lastmod",
         "modified",
         "date",
         "publishdate",
         "pubdate",
         "published"
      ],
      "publishDate": [
         "publishdate",
         "pubdate",
         "published",
         "date"
      ]
   }
}

If you, as an example, have a non-standard date parameter in some of your content, you can override the setting for date:

hugo.
      
frontmatter:
  date:
  - myDate
  - :default

[frontmatter]
  date = ['myDate', ':default']

{
   "frontmatter": {
      "date": [
         "myDate",
         ":default"
      ]
   }
}

The :default is a shortcut to the default settings. The above will set .Date to the date value in myDate if present, if not we will look in date,publishDate, lastmod and pick the first valid date.

In the list to the right, values starting with “:” are date handlers with a special meaning (see below). The others are just names of date parameters (case insensitive) in your front matter configuration. Also note that Hugo have some built-in aliases to the above: lastmod => modified, publishDate => pubdate, published and expiryDate => unpublishdate. With that, as an example, using pubDate as a date in front matter, will, by default, be assigned to .PublishDate.

The special date handlers are:

:fileModTime
Fetches the date from the content file’s last modification timestamp.

An example:

hugo.
      
frontmatter:
  lastmod:
  - lastmod
  - :fileModTime
  - :default

[frontmatter]
  lastmod = ['lastmod', ':fileModTime', ':default']

{
   "frontmatter": {
      "lastmod": [
         "lastmod",
         ":fileModTime",
         ":default"
      ]
   }
}

The above will try first to extract the value for .Lastmod starting with the lastmod front matter parameter, then the content file’s modification timestamp. The last, :default should not be needed here, but Hugo will finally look for a valid date in :git, date and then publishDate.

:filename
Fetches the date from the content file’s file name. For example, 2018-02-22-mypage.md will extract the date 2018-02-22. Also, if slug is not set, mypage will be used as the value for .Slug.

An example:

hugo.
      
frontmatter:
  date:
  - :filename
  - :default

[frontmatter]
  date = [':filename', ':default']

{
   "frontmatter": {
      "date": [
         ":filename",
         ":default"
      ]
   }
}

The above will try first to extract the value for .Date from the file name, then it will look in front matter parameters date, publishDate and lastly lastmod.

:git
This is the Git author date for the last revision of this content file. This will only be set if --enableGitInfo is set or enableGitInfo = true is set in site configuration.

Configure minify

See the tdewolff/minify project page for details.

Default configuration:

hugo.
      
minify:
  disableCSS: false
  disableHTML: false
  disableJS: false
  disableJSON: false
  disableSVG: false
  disableXML: false
  minifyOutput: false
  tdewolff:
    css:
      keepCSS2: true
      precision: 0
    html:
      keepComments: false
      keepConditionalComments: false
      keepDefaultAttrVals: true
      keepDocumentTags: true
      keepEndTags: true
      keepQuotes: false
      keepSpecialComments: true
      keepWhitespace: false
      templateDelims:
      - ""
      - ""
    js:
      keepVarNames: false
      precision: 0
      version: 2022
    json:
      keepNumbers: false
      precision: 0
    svg:
      keepComments: false
      precision: 0
    xml:
      keepWhitespace: false

[minify]
  disableCSS = false
  disableHTML = false
  disableJS = false
  disableJSON = false
  disableSVG = false
  disableXML = false
  minifyOutput = false
  [minify.tdewolff]
    [minify.tdewolff.css]
      keepCSS2 = true
      precision = 0
    [minify.tdewolff.html]
      keepComments = false
      keepConditionalComments = false
      keepDefaultAttrVals = true
      keepDocumentTags = true
      keepEndTags = true
      keepQuotes = false
      keepSpecialComments = true
      keepWhitespace = false
      templateDelims = ['', '']
    [minify.tdewolff.js]
      keepVarNames = false
      precision = 0
      version = 2022
    [minify.tdewolff.json]
      keepNumbers = false
      precision = 0
    [minify.tdewolff.svg]
      keepComments = false
      precision = 0
    [minify.tdewolff.xml]
      keepWhitespace = false

{
   "minify": {
      "disableCSS": false,
      "disableHTML": false,
      "disableJS": false,
      "disableJSON": false,
      "disableSVG": false,
      "disableXML": false,
      "minifyOutput": false,
      "tdewolff": {
         "css": {
            "keepCSS2": true,
            "precision": 0
         },
         "html": {
            "keepComments": false,
            "keepConditionalComments": false,
            "keepDefaultAttrVals": true,
            "keepDocumentTags": true,
            "keepEndTags": true,
            "keepQuotes": false,
            "keepSpecialComments": true,
            "keepWhitespace": false,
            "templateDelims": [
               "",
               ""
            ]
         },
         "js": {
            "keepVarNames": false,
            "precision": 0,
            "version": 2022
         },
         "json": {
            "keepNumbers": false,
            "precision": 0
         },
         "svg": {
            "keepComments": false,
            "precision": 0
         },
         "xml": {
            "keepWhitespace": false
         }
      }
   }
}

Configure file caches

Since Hugo 0.52 you can configure more than just the cacheDir. This is the default configuration:

hugo.
      
caches:
  assets:
    dir: :resourceDir/_gen
    maxAge: -1
  getcsv:
    dir: :cacheDir/:project
    maxAge: -1
  getjson:
    dir: :cacheDir/:project
    maxAge: -1
  getresource:
    dir: :cacheDir/:project
    maxAge: -1
  images:
    dir: :resourceDir/_gen
    maxAge: -1
  modules:
    dir: :cacheDir/modules
    maxAge: -1

[caches]
  [caches.assets]
    dir = ':resourceDir/_gen'
    maxAge = -1
  [caches.getcsv]
    dir = ':cacheDir/:project'
    maxAge = -1
  [caches.getjson]
    dir = ':cacheDir/:project'
    maxAge = -1
  [caches.getresource]
    dir = ':cacheDir/:project'
    maxAge = -1
  [caches.images]
    dir = ':resourceDir/_gen'
    maxAge = -1
  [caches.modules]
    dir = ':cacheDir/modules'
    maxAge = -1

{
   "caches": {
      "assets": {
         "dir": ":resourceDir/_gen",
         "maxAge": -1
      },
      "getcsv": {
         "dir": ":cacheDir/:project",
         "maxAge": -1
      },
      "getjson": {
         "dir": ":cacheDir/:project",
         "maxAge": -1
      },
      "getresource": {
         "dir": ":cacheDir/:project",
         "maxAge": -1
      },
      "images": {
         "dir": ":resourceDir/_gen",
         "maxAge": -1
      },
      "modules": {
         "dir": ":cacheDir/modules",
         "maxAge": -1
      }
   }
}

You can override any of these cache settings in your own hugo.toml.

The keywords explained

cacheDir
(string) See Configure cacheDir.
project
(string) The base directory name of the current Hugo project. This means that, in its default setting, every project will have separated file caches, which means that when you do hugo --gc you will not touch files related to other Hugo projects running on the same PC.
resourceDir
(string) This is the value of the resourceDir configuration option.
maxAge
(string) This is the duration before a cache entry will be evicted, -1 means forever and 0 effectively turns that particular cache off. Uses Go’s time.Duration, so valid values are "10s" (10 seconds), "10m" (10 minutes) and "10h" (10 hours).
dir
(string) The absolute path to where the files for this cache will be stored. Allowed starting placeholders are :cacheDir and :resourceDir (see above).

Configure cacheDir

This is the directory where Hugo by default will store its file caches. See Configure File Caches.

This can be set using the cacheDir config option or via the OS environment variable HUGO_CACHEDIR.

If this is not set, Hugo will use, in order of preference:

  1. If running on Netlify: /opt/build/cache/hugo_cache/. This means that if you run your builds on Netlify, all caches configured with :cacheDir will be saved and restored on the next build. For other CI vendors, please read their documentation. For an CircleCI example, see this configuration.
  2. In a hugo_cache directory below the OS user cache directory as defined by Go’s os.UserCacheDir. On Unix systems, this is $XDG_CACHE_HOME as specified by basedir-spec-latest if non-empty, else $HOME/.cache. On MacOS, this is $HOME/Library/Caches. On Windows, this is%LocalAppData%. On Plan 9, this is $home/lib/cache. New in v0.116.0
  3. In a hugo_cache_$USER directory below the OS temp dir.

If you want to know the current value of cacheDir, you can run hugo config, e.g: hugo config | grep cachedir.

Configure HTTP cache

New in v0.127.0

Note that this configuration is currently only relevant when using the resources.GetRemote function.

The caching in Hugo is layered:

HfdTiyTlnPeacccaaaccchhheee
Dynacache
A in memory LRU cache that gets evicted on changes, Cache Buster matches and in low memory situations.
HTTP Cache
Enables HTTP cache behavior (RFC 9111) for remote resources. This works best for resources with properly set up HTTP cache headers. The HTTP cache uses the file cache to store and serve cached resources.
File Cache
See file cache.

The default HTTP cache disables everything:

hugo.
      
HTTPCache:
  cache:
    for:
      excludes:
      - '**'
      includes: null
  polls:
  - disable: true
    for:
      excludes: null
      includes:
      - '**'
    high: 0s
    low: 0s

[HTTPCache]
  [HTTPCache.cache]
    [HTTPCache.cache.for]
      excludes = ['**']
  [[HTTPCache.polls]]
    disable = true
    high = '0s'
    low = '0s'
    [HTTPCache.polls.for]
      includes = ['**']

{
   "HTTPCache": {
      "cache": {
         "for": {
            "excludes": [
               "**"
            ],
            "includes": null
         }
      },
      "polls": [
         {
            "disable": true,
            "for": {
               "excludes": null,
               "includes": [
                  "**"
               ]
            },
            "high": "0s",
            "low": "0s"
         }
      ]
   }
}
caching
Enabled RFC 9111 cache behavior for a configured set of resources. Stale resources will be refreshed from the file cache even if their configured TTL isn’t reached.
polling
Enables polling for a set of resources. Note that you can enable polling for resources even if HTTP caching is disabled. This setting is only used when in watch mode (e.g. hugo server). When a changed resource is detected, that change triggers a rebuild of pages using that resource.

Configure segments

New in v0.124.0
  • Each segment consists of zero or more exclude filters and zero or more include filters.
  • Each filter consists of one or more field Glob matchers.
  • Each filter in a section (exclude or include) is ORed together, each matcher in a filter is ANDed together.

The fields that can be used in the filters are:

path
The logical page path.
lang
The page language.
kind
The kind of the page.
output
The output format of the page.

It is recommended to put coarse grained filters (e.g. for language and output format) in the excludes section, e.g.:

hugo.
      
segments:
  segment1:
    excludes:
    - lang: n*
    - lang: en
      output: rss
    includes:
    - kind: '{home,term,taxonomy}'
    - path: '{/docs,/docs/**}'

[segments]
  [segments.segment1]
    [[segments.segment1.excludes]]
      lang = 'n*'
    [[segments.segment1.excludes]]
      lang = 'en'
      output = 'rss'
    [[segments.segment1.includes]]
      kind = '{home,term,taxonomy}'
    [[segments.segment1.includes]]
      path = '{/docs,/docs/**}'

{
   "segments": {
      "segment1": {
         "excludes": [
            {
               "lang": "n*"
            },
            {
               "lang": "en",
               "output": "rss"
            }
         ],
         "includes": [
            {
               "kind": "{home,term,taxonomy}"
            },
            {
               "path": "{/docs,/docs/**}"
            }
         ]
      }
   }
}

With the above you can render only the pages in segment1 by configuring the renderSegments or setting the --renderSegments flag:

hugo --renderSegments segment1

Multiple segments can be configured, and the --renderSegments flag can take a comma separated list of segments.

Some use cases for this feature:

  • Splitting builds of big sites.
  • Enable faster builds during development by only rendering a subset of the site.
  • Partial rebuilds, e.g. render the home page and the “news section” every hour, render the entire site once a week.
  • Render only e.g. the JSON output format to push to e.g. a search index.

See also

Last updated: July 5, 2024: hugo doc (a0ca0ab)
Improve this page