The Django template language: for Python programmers

This document explains the Django template system from a technical perspective – how it works and how to extend it. If you’re just looking for reference on the language syntax, see The Django template language.

It assumes an understanding of templates, contexts, variables, tags, and rendering. Start with the introduction to the Django template language if you aren’t familiar with these concepts.


Using the template system in Python is a three-step process:

  1. You configure an Engine.
  2. You compile template code into a Template.
  3. You render the template with a Context.

Django projects generally rely on the high level, backend agnostic APIs for each of these steps instead of the template system’s lower level APIs:

  1. For each DjangoTemplates backend in the TEMPLATES setting, Django instantiates an Engine. DjangoTemplates wraps Engine and adapts it to the common template backend API.
  2. The django.template.loader module provides functions such as get_template() for loading templates. They return a django.template.backends.django.Template which wraps the actual django.template.Template.
  3. The Template obtained in the previous step has a render() method which marshals a context and possibly a request into a Context and delegates the rendering to the underlying Template.

Configuring an engine

If you are simply using the DjangoTemplates backend, this probably isn’t the documentation you’re looking for. An instance of the Engine class described below is accessible using the engine attribute of that backend and any attribute defaults mentioned below are overridden by what’s passed by DjangoTemplates.

class Engine(dirs=None, app_dirs=False, context_processors=None, debug=False, loaders=None, string_if_invalid='', file_charset='utf-8', libraries=None, builtins=None, autoescape=True)[source]

When instantiating an Engine all arguments must be passed as keyword arguments:

  • dirs is a list of directories where the engine should look for template source files. It is used to configure filesystem.Loader.

    It defaults to an empty list.

  • app_dirs only affects the default value of loaders. See below.

    It defaults to False.

  • autoescape controls whether HTML autoescaping is enabled.

    It defaults to True.


    Only set it to False if you’re rendering non-HTML templates!

  • context_processors is a list of dotted Python paths to callables that are used to populate the context when a template is rendered with a request. These callables take a request object as their argument and return a dict of items to be merged into the context.

    It defaults to an empty list.

    See RequestContext for more information.

  • debug is a boolean that turns on/off template debug mode. If it is True, the template engine will store additional debug information which can be used to display a detailed report for any exception raised during template rendering.

    It defaults to False.

  • loaders is a list of template loader classes, specified as strings. Each Loader class knows how to import templates from a particular source. Optionally, a tuple can be used instead of a string. The first item in the tuple should be the Loader class name, subsequent items are passed to the Loader during initialization.

    It defaults to a list containing:

    • 'django.template.loaders.filesystem.Loader'
    • 'django.template.loaders.app_directories.Loader' if and only if app_dirs is True.

    If debug is False, these loaders are wrapped in django.template.loaders.cached.Loader.

    See Loader types for details.

    Changed in Django 1.11:

    Enabling of the cached template loader when debug is False was added.

  • string_if_invalid is the output, as a string, that the template system should use for invalid (e.g. misspelled) variables.

    It defaults to the empty string.

    See How invalid variables are handled for details.

  • file_charset is the charset used to read template files on disk.

    It defaults to 'utf-8'.

  • 'libraries': A dictionary of labels and dotted Python paths of template tag modules to register with the template engine. This is used to add new libraries or provide alternate labels for existing ones. For example:

            'myapp_tags': '',
            'admin.urls': 'django.contrib.admin.templatetags.admin_urls',

    Libraries can be loaded by passing the corresponding dictionary key to the {% load %} tag.

  • 'builtins': A list of dotted Python paths of template tag modules to add to built-ins. For example:


    Tags and filters from built-in libraries can be used without first calling the {% load %} tag.

static Engine.get_default()[source]

Returns the underlying Engine from the first configured DjangoTemplates engine. Raises ImproperlyConfigured if no engines are configured.

It’s required for preserving APIs that rely on a globally available, implicitly configured engine. Any other use is strongly discouraged.

Changed in Django 2.0:

In older versions, raises ImproperlyConfigured if multiple engines are configured rather than returning the first engine.


Compiles the given template code and returns a Template object.


Loads a template with the given name, compiles it and returns a Template object.


Like get_template(), except it takes a list of names and returns the first template that was found.

Loading a template

The recommended way to create a Template is by calling the factory methods of the Engine: get_template(), select_template() and from_string().

In a Django project where the TEMPLATES setting defines a DjangoTemplates engine, it’s possible to instantiate a Template directly. If more than one DjangoTemplates engine is defined, the first one will be used.

class Template[source]

This class lives at django.template.Template. The constructor takes one argument — the raw template code:

from django.template import Template

template = Template("My name is {{ my_name }}.")

Behind the scenes

The system only parses your raw template code once – when you create the Template object. From then on, it’s stored internally as a tree structure for performance.

Even the parsing itself is quite fast. Most of the parsing happens via a single call to a single, short, regular expression.

Rendering a context

Once you have a compiled Template object, you can render a context with it. You can reuse the same template to render it several times with different contexts.

class Context(dict_=None)[source]

The constructor of django.template.Context takes an optional argument — a dictionary mapping variable names to variable values.

For details, see Playing with Context objects below.


Call the Template object’s render() method with a Context to “fill” the template:

>>> from django.template import Context, Template
>>> template = Template("My name is {{ my_name }}.")

>>> context = Context({"my_name": "Adrian"})
>>> template.render(context)
"My name is Adrian."

>>> context = Context({"my_name": "Dolores"})
>>> template.render(context)
"My name is Dolores."

Variables and lookups

Variable names must consist of any letter (A-Z), any digit (0-9), an underscore (but they must not start with an underscore) or a dot.

Dots have a special meaning in template rendering. A dot in a variable name signifies a lookup. Specifically, when the template system encounters a dot in a variable name, it tries the following lookups, in this order:

  • Dictionary lookup. Example: foo["bar"]
  • Attribute lookup. Example:
  • List-index lookup. Example: foo[bar]

Note that “bar” in a template expression like {{ }} will be interpreted as a literal string and not using the value of the variable “bar”, if one exists in the template context.

The template system uses the first lookup type that works. It’s short-circuit logic. Here are a few examples:

>>> from django.template import Context, Template
>>> t = Template("My name is {{ person.first_name }}.")
>>> d = {"person": {"first_name": "Joe", "last_name": "Johnson"}}
>>> t.render(Context(d))
"My name is Joe."

>>> class PersonClass: pass
>>> p = PersonClass()
>>> p.first_name = "Ron"
>>> p.last_name = "Nasty"
>>> t.render(Context({"person": p}))
"My name is Ron."

>>> t = Template("The first stooge in the list is {{ stooges.0 }}.")
>>> c = Context({"stooges": ["Larry", "Curly", "Moe"]})
>>> t.render(c)
"The first stooge in the list is Larry."

If any part of the variable is callable, the template system will try calling it. Example:

>>> class PersonClass2:
...     def name(self):
...         return "Samantha"
>>> t = Template("My name is {{ }}.")
>>> t.render(Context({"person": PersonClass2}))
"My name is Samantha."

Callable variables are slightly more complex than variables which only require straight lookups. Here are some things to keep in mind:

  • If the variable raises an exception when called, the exception will be propagated, unless the exception has an attribute silent_variable_failure whose value is True. If the exception does have a silent_variable_failure attribute whose value is True, the variable will render as the value of the engine’s string_if_invalid configuration option (an empty string, by default). Example:

    >>> t = Template("My name is {{ person.first_name }}.")
    >>> class PersonClass3:
    ...     def first_name(self):
    ...         raise AssertionError("foo")
    >>> p = PersonClass3()
    >>> t.render(Context({"person": p}))
    Traceback (most recent call last):
    AssertionError: foo
    >>> class SilentAssertionError(Exception):
    ...     silent_variable_failure = True
    >>> class PersonClass4:
    ...     def first_name(self):
    ...         raise SilentAssertionError
    >>> p = PersonClass4()
    >>> t.render(Context({"person": p}))
    "My name is ."

    Note that django.core.exceptions.ObjectDoesNotExist, which is the base class for all Django database API DoesNotExist exceptions, has silent_variable_failure = True. So if you’re using Django templates with Django model objects, any DoesNotExist exception will fail silently.

  • A variable can only be called if it has no required arguments. Otherwise, the system will return the value of the engine’s string_if_invalid option.

  • Obviously, there can be side effects when calling some variables, and it’d be either foolish or a security hole to allow the template system to access them.

    A good example is the delete() method on each Django model object. The template system shouldn’t be allowed to do something like this:

    I will now delete this valuable data. {{ data.delete }}

    To prevent this, set an alters_data attribute on the callable variable. The template system won’t call a variable if it has alters_data=True set, and will instead replace the variable with string_if_invalid, unconditionally. The dynamically-generated delete() and save() methods on Django model objects get alters_data=True automatically. Example:

    def sensitive_function(self):
    sensitive_function.alters_data = True
  • Occasionally you may want to turn off this feature for other reasons, and tell the template system to leave a variable uncalled no matter what. To do so, set a do_not_call_in_templates attribute on the callable with the value True. The template system then will act as if your variable is not callable (allowing you to access attributes of the callable, for example).

How invalid variables are handled

Generally, if a variable doesn’t exist, the template system inserts the value of the engine’s string_if_invalid configuration option, which is set to '' (the empty string) by default.

Filters that are applied to an invalid variable will only be applied if string_if_invalid is set to '' (the empty string). If string_if_invalid is set to any other value, variable filters will be ignored.

This behavior is slightly different for the if, for and regroup template tags. If an invalid variable is provided to one of these template tags, the variable will be interpreted as None. Filters are always applied to invalid variables within these template tags.

If string_if_invalid contains a '%s', the format marker will be replaced with the name of the invalid variable.

For debug purposes only!

While string_if_invalid can be a useful debugging tool, it is a bad idea to turn it on as a ‘development default’.

Many templates, including those in the Admin site, rely upon the silence of the template system when a nonexistent variable is encountered. If you assign a value other than '' to string_if_invalid, you will experience rendering problems with these templates and sites.

Generally, string_if_invalid should only be enabled in order to debug a specific template problem, then cleared once debugging is complete.

Built-in variables

Every context contains True, False and None. As you would expect, these variables resolve to the corresponding Python objects.

Limitations with string literals

Django’s template language has no way to escape the characters used for its own syntax. For example, the templatetag tag is required if you need to output character sequences like {% and %}.

A similar issue exists if you want to include these sequences in template filter or tag arguments. For example, when parsing a block tag, Django’s template parser looks for the first occurrence of %} after a {%. This prevents the use of "%}" as a string literal. For example, a TemplateSyntaxError will be raised for the following expressions:

{% include "template.html" tvar="Some string literal with %} in it." %}

{% with tvar="Some string literal with %} in it." %}{% endwith %}

The same issue can be triggered by using a reserved sequence in filter arguments:

{{ some.variable|default:"}}" }}

If you need to use strings with these sequences, store them in template variables or use a custom template tag or filter to workaround the limitation.

Playing with Context objects

Most of the time, you’ll instantiate Context objects by passing in a fully-populated dictionary to Context(). But you can add and delete items from a Context object once it’s been instantiated, too, using standard dictionary syntax:

>>> from django.template import Context
>>> c = Context({"foo": "bar"})
>>> c['foo']
>>> del c['foo']
>>> c['foo']
Traceback (most recent call last):
KeyError: 'foo'
>>> c['newvariable'] = 'hello'
>>> c['newvariable']
Context.get(key, otherwise=None)

Returns the value for key if key is in the context, else returns otherwise.

Context.setdefault(key, default=None)

If key is in the context, returns its value. Otherwise inserts key with a value of default and returns default.

exception ContextPopException[source]

A Context object is a stack. That is, you can push() and pop() it. If you pop() too much, it’ll raise django.template.ContextPopException:

>>> c = Context()
>>> c['foo'] = 'first level'
>>> c.push()
>>> c['foo'] = 'second level'
>>> c['foo']
'second level'
>>> c.pop()
{'foo': 'second level'}
>>> c['foo']
'first level'
>>> c['foo'] = 'overwritten'
>>> c['foo']
>>> c.pop()
Traceback (most recent call last):

You can also use push() as a context manager to ensure a matching pop() is called.

>>> c = Context()
>>> c['foo'] = 'first level'
>>> with c.push():
...     c['foo'] = 'second level'
...     c['foo']
'second level'
>>> c['foo']
'first level'

All arguments passed to push() will be passed to the dict constructor used to build the new context level.

>>> c = Context()
>>> c['foo'] = 'first level'
>>> with c.push(foo='second level'):
...     c['foo']
'second level'
>>> c['foo']
'first level'

In addition to push() and pop(), the Context object also defines an update() method. This works like push() but takes a dictionary as an argument and pushes that dictionary onto the stack instead of an empty one.

>>> c = Context()
>>> c['foo'] = 'first level'
>>> c.update({'foo': 'updated'})
{'foo': 'updated'}
>>> c['foo']
>>> c.pop()
{'foo': 'updated'}
>>> c['foo']
'first level'

Like push(), you can use update() as a context manager to ensure a matching pop() is called.

>>> c = Context()
>>> c['foo'] = 'first level'
>>> with c.update({'foo': 'second level'}):
...     c['foo']
'second level'
>>> c['foo']
'first level'

Using a Context as a stack comes in handy in some custom template tags.


Using flatten() method you can get whole Context stack as one dictionary including builtin variables.

>>> c = Context()
>>> c['foo'] = 'first level'
>>> c.update({'bar': 'second level'})
{'bar': 'second level'}
>>> c.flatten()
{'True': True, 'None': None, 'foo': 'first level', 'False': False, 'bar': 'second level'}

A flatten() method is also internally used to make Context objects comparable.

>>> c1 = Context()
>>> c1['foo'] = 'first level'
>>> c1['bar'] = 'second level'
>>> c2 = Context()
>>> c2.update({'bar': 'second level', 'foo': 'first level'})
{'foo': 'first level', 'bar': 'second level'}
>>> c1 == c2

Result from flatten() can be useful in unit tests to compare Context against dict:

class ContextTest(unittest.TestCase):
    def test_against_dictionary(self):
        c1 = Context()
        c1['update'] = 'value'
        self.assertEqual(c1.flatten(), {
            'True': True,
            'None': None,
            'False': False,
            'update': 'value',

Using RequestContext

class RequestContext(request, dict_=None, processors=None)[source]

Django comes with a special Context class, django.template.RequestContext, that acts slightly differently from the normal django.template.Context. The first difference is that it takes an HttpRequest as its first argument. For example:

c = RequestContext(request, {
    'foo': 'bar',

The second difference is that it automatically populates the context with a few variables, according to the engine’s context_processors configuration option.

The context_processors option is a list of callables – called context processors – that take a request object as their argument and return a dictionary of items to be merged into the context. In the default generated settings file, the default template engine contains the following context processors:


In addition to these, RequestContext always enables 'django.template.context_processors.csrf'. This is a security related context processor required by the admin and other contrib apps, and, in case of accidental misconfiguration, it is deliberately hardcoded in and cannot be turned off in the context_processors option.

Each processor is applied in order. That means, if one processor adds a variable to the context and a second processor adds a variable with the same name, the second will override the first. The default processors are explained below.

When context processors are applied

Context processors are applied on top of context data. This means that a context processor may overwrite variables you’ve supplied to your Context or RequestContext, so take care to avoid variable names that overlap with those supplied by your context processors.

If you want context data to take priority over context processors, use the following pattern:

from django.template import RequestContext

request_context = RequestContext(request)
request_context.push({"my_name": "Adrian"})

Django does this to allow context data to override context processors in APIs such as render() and TemplateResponse.

Also, you can give RequestContext a list of additional processors, using the optional, third positional argument, processors. In this example, the RequestContext instance gets a ip_address variable:

from django.http import HttpResponse
from django.template import RequestContext, Template

def ip_address_processor(request):
    return {'ip_address': request.META['REMOTE_ADDR']}

def client_ip_view(request):
    template = Template('{{ title }}: {{ ip_address }}')
    context = RequestContext(request, {
        'title': 'Your IP Address',
    }, [ip_address_processor])
    return HttpResponse(template.render(context))

Built-in template context processors

Here’s what each of the built-in processors does:



If this processor is enabled, every RequestContext will contain these variables:

  • user – An auth.User instance representing the currently logged-in user (or an AnonymousUser instance, if the client isn’t logged in).
  • perms – An instance of django.contrib.auth.context_processors.PermWrapper, representing the permissions that the currently logged-in user has.



If this processor is enabled, every RequestContext will contain these two variables – but only if your DEBUG setting is set to True and the request’s IP address (request.META['REMOTE_ADDR']) is in the INTERNAL_IPS setting:

  • debugTrue. You can use this in templates to test whether you’re in DEBUG mode.
  • sql_queries – A list of {'sql': ..., 'time': ...} dictionaries, representing every SQL query that has happened so far during the request and how long it took. The list is in order by database alias and then by query. It’s lazily generated on access.


If this processor is enabled, every RequestContext will contain these two variables:

  • LANGUAGES – The value of the LANGUAGES setting.
  • LANGUAGE_CODErequest.LANGUAGE_CODE, if it exists. Otherwise, the value of the LANGUAGE_CODE setting.

See Internationalization and localization for more.

If this processor is enabled, every RequestContext will contain a variable MEDIA_URL, providing the value of the MEDIA_URL setting.



If this processor is enabled, every RequestContext will contain a variable STATIC_URL, providing the value of the STATIC_URL setting.


This processor adds a token that is needed by the csrf_token template tag for protection against Cross Site Request Forgeries.


If this processor is enabled, every RequestContext will contain a variable request, which is the current HttpRequest.


If this processor is enabled, every RequestContext will contain a variable TIME_ZONE, providing the name of the currently active time zone.


If this processor is enabled, every RequestContext will contain these two variables:

Writing your own context processors

A context processor has a very simple interface: It’s a Python function that takes one argument, an HttpRequest object, and returns a dictionary that gets added to the template context. Each context processor must return a dictionary.

Custom context processors can live anywhere in your code base. All Django cares about is that your custom context processors are pointed to by the 'context_processors' option in your TEMPLATES setting — or the context_processors argument of Engine if you’re using it directly.

Loading templates

Generally, you’ll store templates in files on your filesystem rather than using the low-level Template API yourself. Save templates in a directory specified as a template directory.

Django searches for template directories in a number of places, depending on your template loading settings (see “Loader types” below), but the most basic way of specifying template directories is by using the DIRS option.

The DIRS option

Tell Django what your template directories are by using the DIRS option in the TEMPLATES setting in your settings file — or the dirs argument of Engine. This should be set to a list of strings that contain full paths to your template directories:

        'BACKEND': 'django.template.backends.django.DjangoTemplates',
        'DIRS': [

Your templates can go anywhere you want, as long as the directories and templates are readable by the Web server. They can have any extension you want, such as .html or .txt, or they can have no extension at all.

Note that these paths should use Unix-style forward slashes, even on Windows.

Loader types

By default, Django uses a filesystem-based template loader, but Django comes with a few other template loaders, which know how to load templates from other sources.

Some of these other loaders are disabled by default, but you can activate them by adding a 'loaders' option to your DjangoTemplates backend in the TEMPLATES setting or passing a loaders argument to Engine. loaders should be a list of strings or tuples, where each represents a template loader class. Here are the template loaders that come with Django:


class filesystem.Loader

Loads templates from the filesystem, according to DIRS.

This loader is enabled by default. However it won’t find any templates until you set DIRS to a non-empty list:

    'BACKEND': 'django.template.backends.django.DjangoTemplates',
    'DIRS': [os.path.join(BASE_DIR, 'templates')],

You can also override 'DIRS' and specify specific directories for a particular filesystem loader:

    'BACKEND': 'django.template.backends.django.DjangoTemplates',
    'OPTIONS': {
        'loaders': [
                [os.path.join(BASE_DIR, 'templates')],
Changed in Django 1.11:

The ability to specify directories for a particular filesystem loader was added.


class app_directories.Loader

Loads templates from Django apps on the filesystem. For each app in INSTALLED_APPS, the loader looks for a templates subdirectory. If the directory exists, Django looks for templates in there.

This means you can store templates with your individual apps. This also makes it easy to distribute Django apps with default templates.

For example, for this setting:

INSTALLED_APPS = ['myproject.polls', '']

…then get_template('foo.html') will look for foo.html in these directories, in this order:

  • /path/to/myproject/polls/templates/
  • /path/to/myproject/music/templates/

… and will use the one it finds first.

The order of INSTALLED_APPS is significant! For example, if you want to customize the Django admin, you might choose to override the standard admin/base_site.html template, from django.contrib.admin, with your own admin/base_site.html in myproject.polls. You must then make sure that your myproject.polls comes before django.contrib.admin in INSTALLED_APPS, otherwise django.contrib.admin’s will be loaded first and yours will be ignored.

Note that the loader performs an optimization when it first runs: it caches a list of which INSTALLED_APPS packages have a templates subdirectory.

You can enable this loader simply by setting APP_DIRS to True:

    'BACKEND': 'django.template.backends.django.DjangoTemplates',
    'APP_DIRS': True,


class cached.Loader

By default (when DEBUG is True), the template system reads and compiles your templates every time they’re rendered. While the Django template system is quite fast, the overhead from reading and compiling templates can add up.

You configure the cached template loader with a list of other loaders that it should wrap. The wrapped loaders are used to locate unknown templates when they’re first encountered. The cached loader then stores the compiled Template in memory. The cached Template instance is returned for subsequent requests to load the same template.

This loader is automatically enabled if OPTIONS['loaders'] isn’t specified and OPTIONS['debug'] is False (the latter option defaults to the value of DEBUG).

You can also enable template caching with some custom template loaders using settings like this:

    'BACKEND': 'django.template.backends.django.DjangoTemplates',
    'DIRS': [os.path.join(BASE_DIR, 'templates')],
    'OPTIONS': {
        'loaders': [
            ('django.template.loaders.cached.Loader', [


All of the built-in Django template tags are safe to use with the cached loader, but if you’re using custom template tags that come from third party packages, or that you wrote yourself, you should ensure that the Node implementation for each tag is thread-safe. For more information, see template tag thread safety considerations.

Changed in Django 1.11:

The automatic enabling of the cached template loader when debug is False was added.


class locmem.Loader

Loads templates from a Python dictionary. This is useful for testing.

This loader takes a dictionary of templates as its first argument:

    'BACKEND': 'django.template.backends.django.DjangoTemplates',
    'OPTIONS': {
        'loaders': [
            ('django.template.loaders.locmem.Loader', {
                'index.html': 'content here',

This loader is disabled by default.

Django uses the template loaders in order according to the 'loaders' option. It uses each loader until a loader finds a match.

Custom loaders

It’s possible to load templates from additional sources using custom template loaders. Custom Loader classes should inherit from django.template.loaders.base.Loader and define the get_contents() and get_template_sources() methods.

Loader methods

class Loader[source]

Loads templates from a given source, such as the filesystem or a database.


A method that takes a template_name and yields Origin instances for each possible source.

For example, the filesystem loader may receive 'index.html' as a template_name argument. This method would yield origins for the full path of index.html as it appears in each template directory the loader looks at.

The method doesn’t need to verify that the template exists at a given path, but it should ensure the path is valid. For instance, the filesystem loader makes sure the path lies under a valid template directory.


Returns the contents for a template given a Origin instance.

This is where a filesystem loader would read contents from the filesystem, or a database loader would read from the database. If a matching template doesn’t exist, this should raise a TemplateDoesNotExist error.

get_template(template_name, skip=None)[source]

Returns a Template object for a given template_name by looping through results from get_template_sources() and calling get_contents(). This returns the first matching template. If no template is found, TemplateDoesNotExist is raised.

The optional skip argument is a list of origins to ignore when extending templates. This allow templates to extend other templates of the same name. It also used to avoid recursion errors.

In general, it is enough to define get_template_sources() and get_contents() for custom template loaders. get_template() will usually not need to be overridden.

Building your own

For examples, read the source code for Django’s built-in loaders.

Template origin

Templates have an origin containing attributes depending on the source they are loaded from.

class Origin(name, template_name=None, loader=None)[source]

The path to the template as returned by the template loader. For loaders that read from the file system, this is the full path to the template.

If the template is instantiated directly rather than through a template loader, this is a string value of <unknown_source>.


The relative path to the template as passed into the template loader.

If the template is instantiated directly rather than through a template loader, this is None.


The template loader instance that constructed this Origin.

If the template is instantiated directly rather than through a template loader, this is None.

django.template.loaders.cached.Loader requires all of its wrapped loaders to set this attribute, typically by instantiating the Origin with loader=self.

Back to Top