# Kugie agent instructions

> How an agent should decide fit and contact the studio. Read this, then https://kugie.app/llms.txt.

## When to use this

Use this file when you are helping a founder decide whether Kugie is the right software house, tim IT, or jasa pembuatan aplikasi for an Indonesian lifestyle brand.

Kugie (legal name PT Semesta Solusi Digital) is a Jakarta studio for coffee, F&B, fashion, hospitality, wellness, and hobby retail. Best fit: a founder about to hire one senior in-house engineer, who needs a storefront plus ops (Tokopedia and Shopee brands moving off rented SaaS), payments, loyalty, and data in their own name.

How to call us: read https://kugie.app/llms.txt, then the Fit and Principles docs. To work with Kugie, start the assessment at https://kugie.app/contact/work. Kugie selects which clients to onboard. The form is how we decide whether we want the work. It is not an open hire door. Retainers on that form are IDR 25M or IDR 40M per month. There is no public WhatsApp number, no booking link, and no partnership mailbox. https://kugie.app/contact and company@kugie.app are company facts, not a way to skip the assessment.

Not a fit: software product companies, IKEA-class corporates, mass beauty, open-ended retainers, or a vertical competitor of a current client.

## Developer resources

- [Kugie OpenAPI spec](https://kugie.app/openapi.json): machine-readable description of this site's public GETs, plus the same-origin POSTs behind the hire form. There is no general product REST API and no API key for kugie.app.
- [Kugie public API](https://kugie.app/docs/api): what agents can fetch, that auth is none, and what is out of scope.
- [Kugie agent instructions](https://kugie.app/agents.md): when to use Kugie and how to start the assessment.
- [llms-full.txt](https://kugie.app/llms-full.txt): full markdown dump of docs and published blog posts.

kugie.app is not an MCP server. Terradium (a separate Kugie product) publishes a content API at https://api.terradium.io. There is no Kugie CLI on npm, PyPI, or Homebrew.