PostHog Python
https://github.com/posthog/posthog-python- Category
- Developer Tools
- Rank
- No. 423Tools index
- Pricing
- Open Source
- Type
- TOOL
- Builder
- PostHog
- GitHub
- 61 stars
- Latest release
- posthog-v7.47.3
- Date
About
Official PostHog SDK for sending product usage events from Python applications.
What it does
PostHog Python connects application behavior to PostHog while also handling identity, groups, feature flags, contextual tags, exception capture, and AI instrumentation. It can evaluate flags remotely or from locally refreshed definitions, then attach the exact evaluation snapshot to later events. Background queues batch outgoing traffic, while synchronous operation remains available.
Why it's ranked here
This is a substantial application client, not a thin event wrapper. Its strongest case is breadth combined with operational care: separate traffic lanes, bounded flag retries, connection management, local flag evaluation, explicit result types, privacy controls, and event filtering. The repository also maintains framework and OpenFeature documentation, although practical examples live outside the package README.
What's good
Feature flags receive unusually careful treatment. Results can include variants, payloads, reasons, metadata, and exact evaluation snapshots, reducing disagreement between application branching and recorded events. Event delivery supports batching, compression, retries, queue limits, and separate analytics and AI traffic. Context scopes propagate identity, sessions, device identifiers, and tags. Exception capture includes configurable secret masking and variable filtering.
Tradeoffs
Current releases require Python 3.10 or newer. The README deliberately sends readers to external documentation, so the repository alone offers limited onboarding beyond one large interactive example. Local flag evaluation needs a secret credential and refreshed definitions. Remote evaluation still introduces network behavior. Optional AI, telemetry, compression, and framework integrations expand the dependency and configuration surface.
How to use it well
Choose it for Python services that already use PostHog and need analytics, feature decisions, error capture, or AI telemetry under one client. Prefer explicit client instances, scoped context, and saved flag evaluation snapshots. Configure queueing, retries, privacy, and shutdown behavior deliberately. It does not replace the PostHog analytics service or serve as a general observability backend.
Technical notes+
pyproject.toml defines a setuptools package requiring Python 3.10+, core dependencies on requests, backoff, distro, and typing-extensions, plus optional LangChain, Zstandard, and OpenTelemetry groups. posthog/client.py implements distinct analytics and AI queue lanes, background consumers, synchronous capture, context propagation, exception handling, and local or remote flag evaluation. posthog/request.py provides pooled requests sessions, configurable socket options, bounded flag retries, gzip support, host normalization, and fork-safe session resets. posthog/types.py normalizes legacy and current flag responses into typed frozen dataclasses. posthog/utils.py contains serialization and in-memory or Redis-backed flag caches. Tests are packaged under posthog.test and configured through pyproject.toml.
Observed
- License
- MIT
- Primary language
- Python
- Packaging
- PyPI library package named posthog, built with setuptools and wheel
- Interface
- Importable Python library with module-level and explicit client APIs
- Python support
- Requires Python 3.10 or newer
- Platform support
- Classified as operating-system independent
- Core dependencies
- requests, backoff, distro, and typing-extensions
- Repository structure
- Tests live inside the posthog package and use pytest
Read from README.md, Makefile, setup.py, pyproject.toml, example.py, setup_analytics.py, posthog/args.py, posthog/types.py, posthog/utils.py, posthog/client.py, posthog/poller.py, posthog/request.py, posthog/version.py, posthog/__init__.py, posthog/_logging.py.
What it can do
Send product usage events to PostHog
Event data and user properties from Python application → Events tracked in PostHog analytics platform
Track user actions and behaviors
User interaction data and event parameters → Behavioral analytics data in PostHog dashboard
Identify and associate events with users
User identifier and event properties → User-linked event data for analysis
Capture feature flag evaluations
Feature flag keys and user context → Feature usage analytics and A/B testing data
Set user properties and attributes
User identifier and property data → Updated user profiles in PostHog
Group users into cohorts or organizations
Group identifier and user associations → Grouped user data for segmented analytics
Tags
Tech Stack
Media
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.