> For the complete documentation index, see [llms.txt](https://musenai.gitbook.io/musenai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://musenai.gitbook.io/musenai/understanding-musen/our-principles.md). # Our Principles Version: v3.0.0\ Last updated: February 2026 ## What musen Is musen is **AI Radio**. Not a streaming platform. Not a playlist engine. Not a recommender surface. musen is a **continuous, adaptive radio system** that computes audio experiences over time. Its core question is not: *what should the user choose next?*\ But: *what should be happening in this radio stream right now?* Radio is treated as a **living process**, not a sequence of items. ## Core Invariant **Listening is continuous, and value must follow listening time.** Everything in musen derives from this invariant: * product experience * AI architecture * economic model * creator incentives * legal posture * long term moat If a design decision breaks this invariant, it is wrong. ## Product Canon (Non Negotiable) {% stepper %} {% step %} ### Radio First Experience * Effortless listening * Minimal interaction * Long sessions over clicks * Flow over choice musen is designed to be a daily background, not an attention trap. {% endstep %} {% step %} ### Two Broadcast Modes **AI DJ** * Segment based * Replayable * Curated from the musen database * Does not speak to users **Live Radio** * Real time stream * Music plus AI Host * Prompt driven, continuous * Separate lifecycle and accounting These two modes never mix. {% endstep %} {% step %} ### Segments * Editorial radio objects (\~30 minutes) * Containers of tracks, jingles, and speech * Do not modify or derive tracks * Persistent, reusable, shareable Segments are **programming**, not content ownership. {% endstep %} {% endstepper %} ## AI Architecture Doctrine musen is not powered by a single model. It is a **system of systems**. ### Intelligence Stack * **Mothership** — infrastructure and lifecycle * **Admiral** — real time execution * **Fujin** — long horizon radio intelligence * **Editorial Agents** — AI DJ and AI Host Fujin reasons. Editorial agents decide. Execution systems play. No layer collapses into another. ### Memory as Infrastructure * Session based, not click based * Time aware * Intentionally decayed * Long horizon Memory is not a feature. It is infrastructure. ## Economic Doctrine (Project Checkmate) ### Time Based Allocation * Value is generated per listening hour * No global pools * No popularity weighting * No delayed settlement Each listener funds what they listen to. ### Allocation Split * Mechanical value → tracks (by time played) * Performance value → broadcasters and editors Editorial labor is paid. Ownership is not confused with value flow. ### Long Term Objective * Progressive reduction of platform extraction * Bonding at scale * musen becomes infrastructure, not rent seeker ## Creator Doctrine * Creators are not content suppliers * They are broadcasters, editors, and composers * Sustained attention matters more than virality Segments reward programming. Live radios reward curation. No algorithmic gaming. No pooled competition. ## Legal and AI Governance Doctrine * Creator native content first * Explicit provenance and consent * Conservative scope at launch * Proportional compliance by scale musen: * does not train music models * does not ingest external catalogs * does not allow unconsented derivatives * does not blur orchestration and generation Compliance is intrinsic, not retrofitted. ## Category and Moat musen defines **AI Radio** as a new media category. Moat is created by: * continuous listening data * long horizon memory * time native economics * aligned creator incentives * regulatory readiness This system compounds with use. It is hard to imitate without rebuilding everything. ## Long Term Horizon musen is built as **20 to 30 year infrastructure**. musen is not trying to win feature comparisons. It is designed to **own the category of AI Radio**. ## Final Rule If something: * increases choice friction * optimizes for clicks over time * treats audio as discrete items * recentralizes value * weakens creator trust It does not belong in musen. Powered by AI. Tuned to you. Computed over time. --- # 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, and the optional `goal` query parameter: ``` GET https://musenai.gitbook.io/musenai/understanding-musen/our-principles.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal. 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.