Agents now connect to datahashi over OAuth — scoped tokens, no API keys to paste
datahashi
Open console

How it works

From a raw warehouse to a governed API — in four steps.

Connect your warehouse, define your metrics, publish, hand out access. Every change ships from the app — no restarts, no redeploys. Pilot success looks like this: dashboard X and agent Y return the same number.

Getting started Open the console
  1. 01

    Connect your warehouse

    Sign in and point datahashi at Snowflake or BigQuery (also Postgres, MySQL, or DuckDB) with a read-only role. Nothing is copied out, and datahashi never uses more access than you granted.

  2. 02

    Define your metrics

    Write a model in YAML: your business entities, the dimensions you slice by, the measures you aggregate, and the metrics built from them. Validate it in the app before dashboard or agent production — a mistake is caught here, not by a wrong number in the wild.

  3. 03

    Publish it

    Publishing freezes the model and its access rules — who can see which fields, which rows, and a cost ceiling — as one version. Every consumer from then on gets that exact version. Change it later and you publish again; nothing breaks mid-flight.

  4. 04

    Hand out access

    Give a service an API key or an agent an OAuth sign-in, each scoped to what it's allowed to see. It queries over REST or MCP by naming metrics and dimensions. Try any query yourself in the built-in playground. Every call is metered.

Then

One governed model between your data and everything that reads it.

Getting started Security & trust

Connect a warehouse and publish your first model.

Time-to-first-metric is the goal — start with three metrics you'd trust an agent to speak aloud.

Open the console Read getting started