PostHog Ships Self-Driving Agents, Slack App, and Customer-Facing Analytics API.

PostHog announced on LinkedIn three major product updates in its June changelog: Self-driving enters public beta with AI agents that ship product improvements automatically, a new Slack app brings product data queries into threads, and Endpoints reaches general availability as a versioned API for customer-facing analytics.

The releases address different pain points in the product development workflow. Self-driving, previously launched in alpha, now lets teams define guardrails while agents analyze product data and draft code changes. The Slack integration eliminates context switching for data queries. Endpoints removes the need to build custom backends for exposing analytics to end users.

Self-driving mode transforms product analytics into actionable changes. Teams configure rules, and PostHog's agents monitor signals like feature usage, conversion drops, or error rates. When patterns emerge, agents draft pull requests with proposed fixes or improvements.

The public beta expands access beyond the alpha cohort. Teams maintain control through defined boundaries: which metrics to monitor, what types of changes agents can propose, and approval workflows before merging. The system works with existing PostHog tracking, so teams already capturing events can enable Self-driving without additional instrumentation.

PostHog positions this as infrastructure for product teams moving fast. Instead of waiting for weekly reviews to spot issues, agents surface problems and draft solutions as they happen. The guardrails prevent runaway automation while preserving speed.

The PostHog Slack app, also in beta, brings product data into the channels where teams already work. Type @PostHog followed by a question, and the bot responds with data from your PostHog instance directly in the thread.

Queries pull from the same datasets available in the PostHog dashboard: event analytics, user properties, funnels, retention cohorts. Teams can ask about conversion rates, feature adoption, or user segments without opening another tab. The app authenticates against your PostHog workspace, so permissions and data access follow existing role configurations.

This follows the pattern of tools like Notion and Intercom embedding assistants in Slack. For PostHog users, it reduces friction in daily standups or planning conversations when someone asks, "How many users tried the new flow last week?" The answer appears in seconds, sourced from live data.

Endpoints, now generally available, turns any saved PostHog insight or SQL query into a versioned, authenticated API. Teams building customer-facing dashboards or embedded analytics no longer need to build their own API layer, handle caching, or manage authentication separately.

Take a saved funnel or cohort analysis, click "Create Endpoint," and PostHog generates a URL with authentication tokens. That endpoint serves JSON responses with the latest data, versioned so changes to the underlying query don't break existing integrations. Rate limiting and auth are handled by PostHog's infrastructure.

This matters for B2B SaaS products exposing usage data to customers, or consumer apps showing personalized stats. Instead of building a backend service to query PostHog, transform results, and serve them securely, teams call an endpoint. Updates to the underlying query flow through automatically without changing the integration.

PostHog has been investing heavily in infrastructure and developer experience. The company rewrote its feature flag service in Rust, cutting resource use by 95%, and rebuilt its data warehouse to handle both product analytics and data team workloads. Endpoints extends that infrastructure to external use cases, making PostHog data available beyond internal dashboards.

All three features are live now. Self-driving and the Slack app are in beta with signups open. Endpoints is available in all PostHog plans, with pricing based on API call volume beyond included limits.