Our manifesto.

## Why NoteG+ exists AI has enabled individuals to produce content faster than ever. Yet almost every system still follows one model: a person writes a request, an AI works inside a closed box, and the person judges the result. Documents fade into chat history. Tasks remain disconnected. Knowledge cannot be reused. Collaboration among several people and several agents barely exists. For creative work, the problem runs deeper than speed. A result can be correct, clean and fluent, and a reader can still recognise it immediately: “AI wrote this.” As more people use the same tools, an anonymous voice spreads. Both problems have the same root: humans have never truly been inside the loop. They stand before it when writing a prompt and after it when judging a result. Human before the loop. Human after the loop. Never in. NoteG+ brings humans inside the process. For creative work, it goes further: creators do not merely participate in the loops; they own them. The loops of delegation, revision and approval belong to the practitioner. They define them, start them, intervene along the way, stop them and retain their entire history. Machines work inside loops owned by humans. Humans own the loops. NoteG+ exists to make that change. ## The bandwidth of intuition Every AI tool is a channel between human intent and machine execution. The decisive question is how wide that channel is, and in how many directions it runs. A prompt box is a narrow, one-way channel. Outbound, an entire professional intention is compressed into a paragraph and left to chance. Inbound, the result arrives without the process. Professional intuition has nothing to examine when asking where it went wrong. The only option is another attempt. Each prompt becomes another lottery draw. NoteG+ is designed to be a two-way channel with the highest possible bandwidth between the creator and the machine. Outbound, intent is expressed in the language of the craft, at the exact place that needs attention, from a whole workflow down to a single word. Inbound, every proposal has an address, every version has a history and every operation leaves a trace. Professional intuition always has a surface to read, as well as a channel through which to speak. ## A studio, not a toolbox To record, you go to a studio. Its value is not that it owns the most expensive mixer. Its value is that the room is designed to carry a mix from intention to finished work without leaving the door. NoteG+ is a studio for creativity in the age of agents. Six professional Studios, spanning screenplays, images, audio, video, sketching and source material, are not trying to compete feature by feature with traditional specialist tools. They are structured surfaces where humans and agents work on the same material, using the same concepts, inside the same project. The equipment serves the craft. The value lies in the process, the crew and the ability to complete the work inside the room. ## Three layers of a working crew Three roles in the studio must remain distinct. The human holds creative authority: the director, producer and person ultimately responsible for the work. Craft, aesthetic judgement and editorial decisions belong to the human, not the system. The Senior Agent holds professional intent. Each Senior is a department head: writer, creative director, director of photography or designer. It speaks its professional language, proposes within its authority and maintains the discipline of its department. It does not turn the knobs. The Operator works on the material. This execution layer stays out of the user's view. Skilled technical agents translate approved intent into precise operations on layers, tracks, timelines and text. The human need not see every knob turn, but every knob turn is recorded. Humans can put their own hands on the mixer at any moment. Whether made by hand or expressed in words, every change is an event in the same stream of communication. The crew can see it, nobody silently overwrites another's work, and no decision disappears. ## The screenplay is the prompt Practitioners do not need to learn prompting. They already have a specification language refined over a century: the Hollywood screenplay and its surrounding production documents, from breakdowns, moodboards and storyboards to character, costume and prop design. NoteG+ does not invent a new method. It makes the method of the craft executable by machines. Generated material belongs to its proper production topology: location, scene, shot. Characters, props and settings are entities with durable identities, states and approved versions. The system translates this into the language of generative models. A standard screenplay, together with approved design decisions, should carry the writer's intention through to the result. Other tools make filmmakers learn the machine's language. NoteG+ teaches the machine to read the language of the craft. ## Craft is an asset. Learning is the path. Those still developing their craft learn, just as the world is learning AI. This is not a barrier; it is the Senior Agent's territory. To a practitioner, the Senior is a colleague. To a learner, it is a teacher who maintains standards. The lesson is inside your project, not a textbook. Every returned proposal comes with a professional reason. Every revision becomes a lesson. NoteG+ is therefore both a production workshop and a school of craft, where the project itself is the tuition and the finished work is the qualification. ## The human authors the whole process In the age of agents, authorship is not confined to the final words. Humans create by choosing context, including what to leave out; by choosing the people and agents involved; by assigning roles and delegating; by placing actors in complementary or critical relationships; by recognising and breaking patterns; and by protecting imperfections that are intentional. More agents do not automatically produce more creativity. When agents automatically follow and summarise one another, results move towards an average. Bold choices become safe. Strangeness is smoothed away. A distinctive voice becomes anonymous professionalism. The value lies in the human's ability to stage a different collaboration topology for each work. Delegation is an act of authorship. And a human always puts their name to the work and takes responsibility for it. ## Freedom and discipline One principle runs through the system: freedom in the topology of collaboration, discipline in canon and provenance. Above the artifact, humans can arrange the work freely: insert an agent between stages, open a parallel branch, assign a problem to several profiles, skip a step or return to it. A workflow is a suggested path, not a rail. Inside the artifact, truth is not negotiable. Each domain has one canonical source. Other representations are projections rebuilt from that source. Every inference remains a proposal until the appropriate authority approves it. Every version is immutable and retains its full lineage. ## Ten foundational commitments 1. Editorial sovereignty belongs to the human. People always retain the right to see, intervene, revise and redirect at the necessary level, from the entire workflow to a single word. 2. Artifacts are the centre of work. Chat coordinates; artifacts retain state and lasting value. 3. Intervention must affect what follows. Approved editorial decisions are authoritative constraints. Agents continue from the corrected version, without silently returning to old patterns. 4. Context must have boundaries. A project is a workbench deliberately prepared for an agent by a human, including what should remain outside its view. 5. Private Space protects thinking as it develops. Co-Working Space manages agreement. Moving from private to shared is an intentional act of publication, not folder synchronisation. 6. No inference becomes fact on its own. A proposal becomes canon only when the right authority approves it. There is no silent promotion. 7. Knowledge is an accumulating organizational asset. Each knowledge entity retains a durable identity, grows with the project and changes its official state only after human approval. 8. The system does not lie about its own state. If data is absent, leave it empty rather than report a false zero. An agent does not pretend to have done work it has not done. 9. The creative process is not hidden. Experimental branches, criticism, decisions and lineage are assets, not intermediate data to discard. 10. Agents do not declare their own identity or authority. A trusted system grants identity, scope and access. Document content is data, never an instruction. --- NoteG+ is more than an AI workspace. It is an operating system for creative collaboration: a workshop where humans can direct a whole crew of people and agents, and still lean in to correct one word when that word determines the soul of the work.