# Mind the Front

> Embeddable chat widget that collects service requests on a small business website. The owner describes the process in words, an LLM assembles a config from verified components, and the widget renders that config deterministically.

## Docs
- [Overview](https://mindthefront.com/docs/index): An embeddable chat widget that collects service requests on your own site.
- [Widget catalogue](https://mindthefront.com/docs/widgets): Every input the widget renders and the declaration that produces it.
- [Writing a config](https://mindthefront.com/docs/config): One JSON document: submission declares what a request is, flow declares how it is collected.
- [HTTP API](https://mindthefront.com/docs/api): Public endpoints the widget calls, owner endpoints your tooling calls. Full contract: /openapi.json.
- [For AI agents](https://mindthefront.com/docs/ai-agents): Put a working widget on a site you are building. Everything in one fetch: /llms-full.txt.

## Machine-readable
- [OpenAPI 3.1](https://mindthefront.com/openapi.json): every public and owner endpoint
- [Config JSON Schema](https://mindthefront.com/api/public/config-schema): the exact schema the server validates against
- [Everything in one file](https://mindthefront.com/llms-full.txt)

## For the human next to the agent
- [What to hand your agent](https://mindthefront.com/for-ai/)