Product Scope

At the current stage, Myros is packaged as an open-source project.

NOW LIVE! Launch Open Beta here | Myros Framework here

The initial release focuses on:

Area	Current Scope
Deployment model	Self-hosted open-source framework
Primary interface	Telegram
Knowledge source	Website crawling from admin-approved URLs
AI layer	OpenAI-backed responses with configurable model settings
Persistence	PostgreSQL for projects and conversation history
Customization	Editable personality and agent behavior
Administration	Telegram-native admin controls
Access model	Project-specific agents with user and admin modes

Myros is intentionally starting with a practical and usable foundation rather than an abstract promise.

The first goal is to give builders a working base for launching real-time knowledge agents.

Open-Source Philosophy

Myros is being launched as an open-source project.

This is not only a distribution choice. It is part of the product philosophy.

AI agents will increasingly become the interface through which people access knowledge, communities, products, and services. If these agents are closed, centralized, and controlled by black-box systems, builders lose control over how their own knowledge is represented.

Myros takes a different path.

Builders should be able to see how their agents work. Communities should be able to trust how answers are generated. Developers should be able to extend the system. Projects should be able to self-host their own intelligence layer. Users should not be forced into one centralized platform.

Open source gives Myros the foundation for transparency, extensibility, and community-led evolution.

---

Why Open Source Matters for AI Agents

AI agents are not ordinary applications.

They speak on behalf of projects. They represent brands. They shape user perception. They influence onboarding, support, education, and trust.

For that reason, projects need control.

Open-Source Principle	Myros Application
Transparency	Builders can inspect how Myros retrieves and processes knowledge
Self-hosting	Projects can operate their own agent infrastructure
Extensibility	Developers can build plugins, integrations, and custom modules
Community contribution	Improvements can come from users and builders
Portability	Projects are not locked into one centralized service
Trust	The agent’s behavior can be audited and modified

Myros is designed for builders who want ownership over their AI layer.

---

Built to Be Forked, Modified, and Extended

Myros is not meant to be a closed product that users can only consume.

It is meant to be a base layer that builders can adapt.

Developers may extend Myros by adding:

Extension Area	Example
New interfaces	Discord, Slack, web chat, in-app agents
New data sources	PDFs, GitHub repos, Notion, Google Drive, databases
New retrieval systems	Embeddings, vector search, hybrid search
New memory systems	Long-term user memory, project memory, role-based memory
New admin tools	Dashboards, analytics, permission systems
New agent personalities	Technical support, lorekeeper, community guide, analyst

The larger opportunity is an ecosystem of community-built agent modules.