Alpha · Early access

Game analytics you can ask in plain English.

Ask your AI coding agent about retention, funnels and player segments. LensGlint turns the answer into charts your whole team can use.

The problem

Generic BI wasn't built for games.

Every question is a SQL ticket

Producers wait on a data person for answers they need today.

Game analysis is hand-built

Retention, funnels and segments have to be stitched together by hand in general-purpose BI.

Parameters get buried

Event parameter limits push fields into JSON blobs — and then nobody can analyse them.

How it works

From question to dashboard without a SQL ticket.

  1. 1

    Ask

    Ask in your agent

    Use /lens-analyze in Claude Code — or call the same tools from any MCP client.

  2. 2

    Query

    Run it on your events

    LensGlint runs a structured query or read-only SQL against your event data.

  3. 3

    Save

    Keep the answer

    Results become cards with the chart, the interpretation and the query behind it.

  4. 4

    Share

    Share with the team

    Cards come together as dashboards your team opens in the browser.

Analysis models

The four questions every game team asks.

Events

Any event by any dimension, with counts, uniques, sums and averages over time.

Retention

Cohort matrices from any start event to any return event, by day.

Funnels

Step-by-step conversion within a time window you choose.

Segments

Find players by what they did — or didn't do.

For the whole team

Answers that outlive the conversation.

  • Cards you can trust

    Every card keeps its chart, its interpretation and the query that produced it.

  • An inbox for AI output

    New results land in an inbox first; you decide which dashboard they belong on.

  • Personal and team spaces

    Draft in your own space, then publish to the team when it is ready.

Your events, understood

A semantic layer built for game telemetry.

Event and property catalogue

One place for what every event and property means.

Virtual properties

Declare fields packed into JSON parameters as virtual properties and analyse them like any other.

Readable values

Map internal codes to display names, so charts read like the game.

Built to be trusted

Your data stays yours.

Bring your own AI

LensGlint never calls an AI provider. Analysis happens in your own agent.

Read-only SQL

Explore freely: queries from agents and people are read-only.

Isolated projects

Each game's data is kept separate from every other project.

Org, studio and project roles

Grant access at the level that matches how your studio works.

Your infrastructure

Designed to run in your own cloud.

Data source

v1 works with the Firebase Analytics export to BigQuery.

In internal alpha on live mobile games.

FAQ

Questions, answered.

Which AI tools does it work with?

Claude Code is supported first-class, including the /lens-* slash commands. Any client that supports stdio MCP servers — such as Cursor, Codex or Windsurf — can call the same tools.

Does LensGlint send my data to an AI provider?

No. LensGlint itself never calls an AI provider. Analysis happens in your own agent, under the terms of the AI provider you already use; the agent sees the query results you ask it to work with.

What data sources are supported?

v1 works with the Firebase Analytics export to BigQuery.

When can we use it?

LensGlint is in internal alpha. We will invite studios from the waitlist in small groups.

How much will it cost?

Pricing will be shared with early-access studios.

Early access

Get early access to LensGlint.

Tell us about your studio, your games, your analytics stack and your team. We'll be in touch as we open up the alpha.

Join the waitlist