Datost Documentation archiveCurrent E1O project ↗
Historical documentation · May 7, 2026

This snapshot describes Datost’s former Slack-native analytics product. It is retained for reference; setup instructions, service availability and external links may have changed. It does not describe E1O glasses.

PostHog

Connect PostHog to Datost to query product analytics, session replays, feature flags, and experiments from Slack.

Once connected, your team can ask Datost questions about product usage, funnels, and experiments in plain English and have them answered against your live PostHog project.

What you get

  • Ask questions like @Datost how many signups did we have last week by source? in Slack and get answers pulled from PostHog events.
  • Query events, persons, cohorts, insights, dashboards, feature flags, experiments, surveys, and session replays.
  • Run arbitrary HogQL against your project when a saved insight does not exist.
  • Inspect LLM costs, error tracking issues, and property definitions alongside the rest of your product data.

Prerequisites

Before connecting, make sure you have:

  • A PostHog Cloud account (US or EU) or a self-hosted instance reachable over HTTPS.
  • Admin access to at least one PostHog project you want Datost to read.
  • The ability to complete an OAuth flow on behalf of your org.

Connect PostHog

Open the admin panel

In the Datost web app, go to Data Sources and click Add data source.

Pick PostHog

Select PostHog from the list of business tools. Enter your PostHog API host if you are on EU Cloud or self-hosted, otherwise leave the default (https://us.posthog.com).

Sign in with PostHog

Click Connect with PostHog. You will be redirected to PostHog to authorize Datost with the read scope. After you approve, PostHog redirects back to Datost and the access and refresh tokens are stored encrypted against your connection.

Project auto-discovery

On successful OAuth, Datost calls GET /api/projects/ and automatically sets the first project on the connection. If you have multiple projects, you can change the project ID from the connection settings afterwards.

Test and save

Datost runs a test query against the selected project to confirm credentials work, then saves the connection. Your team can start asking questions immediately.

What Datost can query

Datost exposes the following PostHog surfaces to the agent:

  • Events — query by event name, time range, distinct ID, and properties using HogQL.
  • Persons and cohorts — look up users by distinct ID or email, and read cohort membership.
  • Insights and dashboards — list and read saved insights, including trends, funnels, and retention.
  • Feature flags and experiments — list flags and experiments, and compute experiment results from exposure queries.
  • Surveys — list surveys and fetch response stats.
  • Session replays — search replays by person or time range and render them inline in Slack.
  • Error tracking and LLM costs — list issues, read details, and roll up LLM spend over a given window.
  • Schema discovery — list event and property definitions so Datost knows what it can ask about.

Limitations

  • Datost has read-only access. It cannot capture events, toggle feature flags, or edit insights.
  • OAuth tokens are refreshed automatically one minute before expiry. If a refresh fails, the connection is marked unhealthy and you will need to reconnect.
  • Self-hosted PostHog must be reachable over HTTPS from Datost's backend. Private IPs, IPv6 literals, and cloud metadata endpoints are rejected.
  • Timestamps in HogQL queries are converted to YYYY-MM-DD HH:MM:SS format since ClickHouse does not accept ISO-8601 with T and Z.
  • A connection is scoped to one PostHog project. Add a separate Datost connection for each project you want to query.