# What is MIRARI?

MIRARI is a **self-hostable web app** that turns any OpenAI-compatible LLM endpoint into a *persistent, self-improving agent*. It is built around three ideas that most chat tools ignore:

1. **Memory is a graph, not a transcript.** Conversations are ephemeral; *facts* are not. MIRARI lets you promote anything worth remembering into a **Memory Atlas** — a strength-weighted graph of nodes and edges that any future session can pull from.
2. **Skills should be versioned, not improvised.** Every clever prompt, workflow, or rubric the agent invents can be saved as a **Skill** with a status (`suggested → accepted → testing → promoted → archived`) and a version number. Skills become reusable instruments, not throwaway tokens.
3. **Improvement is a ritual, not a vibe.** **Mirror Mode** takes raw conversation logs and uses an analysis pass to produce a structured report — strengths, weaknesses, recommendations, and a *suggested system prompt* you can A/B test against the current one.

***

## Who it's for

* **Builders running Hermes locally** who want a real UI instead of a curl loop.
* **Power users of OpenRouter, Nous, or any OpenAI-style endpoint** who want one dashboard across providers.
* **Anyone tired of starting from scratch every chat** and watching the same context evaporate every Monday morning.

***

## What it is *not*

* Not a hosted LLM. MIRARI calls *your* model — your endpoint, your keys, your costs.
* Not a vector DB you have to operate. The graph lives in your backend and Just Works.
* Not opinionated about which model you use. Hermes is the first-class preset; OpenRouter, Nous, and arbitrary base URLs are equally supported.

***

## The Six Rites

MIRARI's feature set is organized as six "rites":

| Rite        | Page      | What it does                                             |
| ----------- | --------- | -------------------------------------------------------- |
| ☿ Bind      | Settings  | Connect any OpenAI-compatible endpoint                   |
| ◈ Console   | Console   | Chat with persistent threads, promote messages to memory |
| ✶ Atlas     | Memory    | Inspect and curate the memory graph                      |
| ⚚ Forge     | Skills    | Suggest, score, version, and promote skills              |
| ◉ Mirror    | Mirror    | Analyze logs → strengths, weaknesses, suggested prompt   |
| ⌘ Templates | Templates | Share / import skills as community templates             |

Plus an **Observatory** (dashboard) and a **Tasks** workboard.


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://entermirari.gitbook.io/entermirari-docs/getting-started/what-is-mirari.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.