OpenReplay
Self-hostable session replay and product analytics. Your users behavioural data stays on your own infrastructure instead of going to a third party.
Strengths
- Full session replay with console, network requests, JS errors and performance metrics on a single timeline
- Genuine self-hosting: no user data leaves your infrastructure, no session quota to buy back
- Fine-grained masking controls, field by field, so sensitive input is never captured
- Co-browsing lets you take over a user session live during a support call
Limitations
- The install is a real stack: PostgreSQL, ClickHouse, Kafka and S3-compatible object storage. The documented minimum is 2 vCPUs, 8 GB RAM and 50 GB disk, and that is the evaluation threshold, not the production one.
- Docker Compose is for single-machine evaluation. Going to production means Kubernetes, and therefore someone who can tune ClickHouse when session volume climbs.
- Product analytics covers funnels, journeys and heatmaps, but not the feature flags and A/B tests PostHog bundles. If you gate releases with flags, you keep two tools.
- The core is AGPL v3: embedding the code in a distributed product has legal consequences worth clearing before you start.
Best for
- Products handling health, banking or HR data, where shipping sessions to a US SaaS is a non-starter
- High-traffic teams whose analytics bill already exceeds the cost of a dedicated server
- Product support teams needing to take over a live session rather than asking for screenshots