For the work that matters.
The agent can change. The project remains. Make any agent your auteur. You stay the creative authority: every agent works inside your project, under your approvals. Auteur Intelligence™ turns the AI models and tools you already use into one persistent, governed studio. It carries your approved decisions across story, production, film, software, games, and launch, while specialized agents do the work, check each other, and leave a receipt for every step.
See Auteur work Apply for the Founding Beta
Bring your subscriptions, APIs, creative applications, and local compute.
Built to ride the curve, not fight it.
Frontier models will keep getting stronger. Inside Auteur, that does not erase the project or the workflow. It improves the seats available to them. A better model is a better hire in your studio, not a new studio.
Your canon, approvals, tools, evidence, and production history stay independent of any one model or interface. Models from more than one company work the same project and check each other's results, which is also why smaller, cheaper models can carry governed work that once needed a frontier seat. The agent can change. The project remains.
Seats do not only report to you; they report to each other. Every seat, from any company or your own hardware, reads and writes the same project inbox, so a note from one model lands in front of another and the whole exchange stays in project memory with a receipt. Direct seat-to-seat messaging across harnesses is being built now on the coordination layer the engine already uses for its own campaigns.
Complex work deserves more than a transcript.
A chat session is a good place to start a thought and a bad place to keep one. It forgets. It cannot tell an approved decision from a passing suggestion. It has no record of what it spent or what it changed. Auteur replaces the disappearing transcript with three durable things.
Persistent project intelligence
Approved canon, research, decisions, sources, and production history survive the session and the model change. Approved truth stays separate from suggestions.
Governed execution
Complex goals become durable workflows with scoped tools, budgets, permissions, approval gates, retries, and recovery. An estimate before, a receipt after.
Independent verification
Work is accepted because the evidence agrees, not because the agent said it was finished. The checker is never the builder.
Find your door.
However you make things, on the page, on set, in the cut, or in code, start at the door that fits.
- Writers and story teams finish long-form work in a project that keeps its own canon and editorial history. → Auteur Story
- Producers and departments turn a script or book into evidence-linked breakdowns a human approves. → Auteur Prep
- Filmmakers and finishers carry an approved story through shots, generation, assembly, and finishing. → Auteur Film
- Developers and technical teams run the models they already have through governed campaigns that verify what changed. → Auteur Build
All ten products, grouped, with current availability
Dashboard optional. Depth optional. Both yours.
Run the whole studio from Claude, Codex, or Grok in plain language and never open the dashboard, or open the Auteur interface and go lane by lane, knob by knob. Guided automation for less technical users, granular inspection and control for experts. Same engine, your altitude.
The engine is ours. The fuel is yours.
Auteur is an amplification and accountability layer, not a replacement for the stack you have chosen. Bring your own AI harness, model subscriptions, API keys, pay-as-you-go accounts, creative applications, or local open-source models on your own GPU. Auteur adds the persistent project, governed campaigns, verification, and receipts around them, and you can change AI interfaces without losing the project, the approvals, or the evidence.
We are model agnostic and we are not a model provider: no resold tokens, no meter. What you need depends on the posture you choose.
What you bring, what we provide · Designed for the agents you use
Outcomes you can audit.
The industry's own safety reports describe agents that misled their evaluators or slipped their boundaries. Auteur's answer is structural rather than hopeful: guardrails, spending limits, private workspaces, human gates, and review by an AI from a different company, so no model verifies its own work.
- Approved production truth is human-reviewed and versioned, and approved canon stays separate from automated suggestions.
- Auteur derives commits, changed paths, tests, and mismatches independently of what an agent reports.
- Agent work runs in private, sandboxed workspaces and is verified before anything is integrated.
Built by people who deliver real work.
Auteur Intelligence™ was built by working filmmakers, writers, editors, and builders who needed AI to survive real projects, not just produce convincing demos. Its founder has spent more than 25 years in the film industry as a writer, director, and editor of award-winning features and documentaries. The gates, the lanes, the receipts, and the taste in every workflow come from real sets, real edit bays, and real deliveries.
Our own engine. On purpose.
Auteur is not built on anyone's harness; those connect as seats. The Algernon Engine underneath is proprietary, designed from the ground up as the moat around your creative work: memory, eval, workshop, receipts, and authority are ours, so they answer to you and to no one's roadmap.
Bring your next real project.
Every auteur needs a studio. We are opening ours to the first 100 founding members: writers, filmmakers, producers, developers, and technical creators who want to shape the studio around demanding real-world work. Seats, when offered, are paid, and founding members keep their rate when prices rise. Every seat includes our course community on Skool, led by professionals who make films, write books, and build software for a living. The questionnaire takes about ten minutes.
