Product health

Write specs against the errors your users actually hit

In sprint planning the most confident voice usually wins, numbers or no numbers. Connect PostHog and Spectr keeps a read-only picture of what is actually happening in production: the errors users hit, the feedback they left in surveys, and which flags are live. It then quotes those numbers, with the date, in the next spec it writes. A claim about what customers keep hitting becomes something you can check.

How it connects

A read-only Personal API Key plus your PostHog region. The connect dialog names the three scopes it needs so you can issue a narrow key rather than an all-access one.

Before you start

PostHog Cloud (US or EU) only. Self-hosted instances are deliberately not supported.

What Spectr reads from PostHog

Errors your users hit: Grouped error data with occurrence counts and the affected service, read through PostHog's query interface. Aggregates only.
Survey feedback: Survey responses, with emails and identifiers scrubbed out before anything is stored.
Feature-flag rollout state: Which flags exist and how far each is rolled out - so a spec can say whether the thing being discussed is even live yet.

A person approves everything before it is published

Spectr generates specs and stories, then stops. Every story goes to a review page where you approve it, edit it inline, or reject it - individually or in bulk. Only approved stories reach an issue tracker. There is no auto-publish setting to switch on, because a backlog nobody read is not a saving.

Setting it up

In PostHog, create a Personal API Key with project read, query read, and feature-flag read scopes.
In Spectr, open the project, go to Integrations, choose PostHog, and enter the key plus your region.
Pick the PostHog project to track. Product health is one-to-one.
Spectr runs an initial sync. Open the Product Health page, or generate a spec and see the dated evidence in its context.

Frequently asked questions

Does Spectr store personal data from PostHog?

No. Only aggregate counts are stored, and every verbatim is scrubbed of emails and identifiers before it is written. This is a hard rule, not a setting.

Why do I need to pick a region?

PostHog Cloud runs separate US and EU deployments, and a project id only means something on the deployment it came from. Spectr resolves the host from a fixed list rather than accepting an arbitrary one, which is also what keeps a user-supplied URL from becoming a way to make our servers call somewhere they should not.