A powerful Elixir SDK for PostHog
- Analytics and Feature Flags support
- Error tracking support
- Powerful process-based context propagation
- Asynchronous event sending with built-in batching
- Overridable HTTP client
- Support for multiple PostHog projects
Add PostHog to your dependencies:
def deps do
[
{:posthog, "~> 2.0"}
]
endConfigure the PostHog application environment:
config :posthog,
enable: true,
enable_error_tracking: true,
api_host: "https://us.i.posthog.com", # Or `https://eu.i.posthog.com` or your self-hosted PostHog instance URL
api_key: "phc_my_api_key",
in_app_otp_apps: [:my_app]For test environment, you want to enable test_mode:
config :posthog,
test_mode: trueOptionally, enable Plug integration.
You're all set! 🎉 For more information on configuration, check the PostHog.Config module
documentation and the advanced configuration guide.
To capture an event, use PostHog.capture/2:
iex> PostHog.capture("user_signed_up", %{distinct_id: "distinct_id_of_the_user"})You can pass additional properties in the last argument:
iex> PostHog.capture("user_signed_up", %{
distinct_id: "distinct_id_of_the_user",
login_type: "email",
is_free_trial: true
})PostHog.capture/2 is very powerful and allows you to send events that have
special meaning. For example:
iex> PostHog.capture("$create_alias", %{distinct_id: "frontend_id", alias: "backend_id"})iex> PostHog.capture("$groupidentify", %{
distinct_id: "static_string_used_for_all_group_events",
"$group_type": "company",
"$group_key": "company_id_in_your_db"
})Carrying distinct_id around all the time might not be the most convenient
approach, so PostHog lets you store it and other properties in a context.
The context is stored in the Logger metadata, and PostHog will automatically
attach these properties to any events you capture with PostHog.capture/3, as long as they
happen in the same process.
iex> PostHog.set_context(%{distinct_id: "distinct_id_of_the_user"})
iex> PostHog.capture("page_opened")You can scope context by event name. In this case, it will only be attached to a specific event:
iex> PostHog.set_event_context("sensitive_event", %{"$process_person_profile": false})You can always inspect the context:
iex> PostHog.get_context()
%{distinct_id: "distinct_id_of_the_user"}
iex> PostHog.get_event_context("sensitive_event")
%{distinct_id: "distinct_id_of_the_user", "$process_person_profile": false}Evaluate feature flags once for a user with PostHog.FeatureFlags.evaluate_flags/1,
then read values from the returned snapshot.
# Boolean feature flag
{:ok, snapshot} = PostHog.FeatureFlags.evaluate_flags("user123")
if PostHog.FeatureFlags.Evaluations.enabled?(snapshot, "new-dashboard") do
# Do something differently for this user
end
# Multivariate feature flag
case PostHog.FeatureFlags.Evaluations.get_flag(snapshot, "checkout-flow") do
"variant-a" -> :variant_a
true -> :enabled_boolean_flag
false -> :disabled
nil -> :not_returned
end
# Optional payload
payload = PostHog.FeatureFlags.Evaluations.get_flag_payload(snapshot, "checkout-flow")get_flag/2 returns the variant string for multivariate flags, true for enabled
boolean flags, false for disabled flags, and nil when the flag was not returned
by the evaluation.
If you want to break down or filter captured events by feature flag value, put the same snapshot in the process context before capturing events:
{:ok, snapshot} = PostHog.FeatureFlags.evaluate_flags("user123")
if PostHog.FeatureFlags.Evaluations.enabled?(snapshot, "new-dashboard") do
# Do something differently for this user
end
PostHog.FeatureFlags.set_in_context(snapshot)
PostHog.capture("page_viewed", %{distinct_id: "user123"})This attaches $feature/<flag-key> properties and $active_feature_flags without
making another /flags request. To reduce event property bloat, filter the
snapshot first:
# Attach only flags accessed with enabled?/2 or get_flag/2
PostHog.FeatureFlags.set_in_context(
PostHog.FeatureFlags.Evaluations.only_accessed(snapshot)
)
# Or attach only specific flags
PostHog.FeatureFlags.set_in_context(
PostHog.FeatureFlags.Evaluations.only(snapshot, ["checkout-flow", "new-dashboard"])
)By default, evaluate_flags/1 evaluates every flag for the user. If you only need
a few flags, pass :flag_keys to request only those flags:
{:ok, snapshot} =
PostHog.FeatureFlags.evaluate_flags(%{
distinct_id: "user123",
flag_keys: ["checkout-flow", "new-dashboard"]
})
PostHog.FeatureFlags.check/2,PostHog.FeatureFlags.check!/2,PostHog.FeatureFlags.get_feature_flag_result/2, andPostHog.FeatureFlags.get_feature_flag_result!/2still work during the migration period, but preferevaluate_flags/1for new code.
Error Tracking is enabled by default.
You can always disable it by setting enable_error_tracking to false:
config :posthog, enable_error_tracking: falseThe SDK uses Req under the hood with gzip compression and transient retry enabled by default. You can swap in your own HTTP client module to change any of this behaviour — disable compression, add custom headers, attach telemetry, or use a completely different HTTP library.
Set the api_client_module config option to a module that implements the
PostHog.API.Client behaviour:
config :posthog, api_client_module: MyApp.PostHogClientThe simplest approach is to wrap the default client and override only what you need:
defmodule MyApp.PostHogClient do
@behaviour PostHog.API.Client
@impl true
def client(api_key, api_host) do
default = PostHog.API.Client.client(api_key, api_host)
# Disable gzip compression
custom = Req.merge(default.client, compress_body: false)
%{default | client: custom}
end
@impl true
defdelegate request(client, method, url, opts), to: PostHog.API.Client
endSee PostHog.API.Client docs for more examples, including adding custom headers
and using a different HTTP library entirely.
If your app works with multiple PostHog projects, PostHog can accommodate you. For setup instructions, consult the advanced configuration guide.
See CONTRIBUTING.md for local setup, integration test, and pull request guidelines.