How Subtxt Uses Evaluations to Build Smarter AI for Better Stories

Smarter AI, sharper stories, constant evolution

At Narrative First, we believe storytelling is as much an art as it is a science. That’s why we’re constantly refining the tools we offer to help you bring your stories to life. A big part of this process is evaluating and improving the AI models that power features like BrainstormingAI. These evaluations don’t just measure success—they guide us toward smarter, more insightful storytelling tools.

Recently, we compared two of our AI models—the Standard model and the Advanced model—to see how they measure up in delivering narrative insights. These evaluations don’t just highlight what’s working; they reveal where we can dig deeper to improve. Here’s what we found—and how it’s shaping the future of Subtxt.

Testing What Matters Most

Every story has a structure, a progression, and a reason to exist. To ensure our AI understands these elements, we evaluate it across four key areas:

1. Method as a Source of Conflict: Can the AI recognize how a character’s approach to solving problems creates tension in the story?
2. Understanding the Definition of the Method: Does the AI grasp what the Method really means and how it operates in context?
3. Avoiding Conflict Pitfalls: Can it avoid common storytelling traps, like substituting vague conflicts ("struggling to," "threatening to") for meaningful tension?
4. Development of Story Context: Does it help build and deepen the narrative as the story progresses?

These criteria ensure that Subtxt’s AI isn’t just generating ideas but genuinely helping you create narratives that resonate.

The Results: Advanced Model Outperforms, But There’s More to Learn

Our evaluations revealed that the Advanced model consistently scored higher than the Standard model across most criteria. For example:

• In Avoiding Conflict, the Advanced model scored 92%, outperforming the Standard model’s 85%. This means it’s better at identifying direct and meaningful sources of tension.
• In Understanding the Definition of the Method, the Advanced model nearly aced it with 99%, compared to the Standard model’s 97%.

But even with these wins, there’s always room for growth. When it came to the Development of Story Context, the Advanced model scored slightly lower (81%) than the Standard model (82%). This told us something important: while the Advanced model excels at identifying and defining conflict, there’s still work to be done in helping it guide users through the nuanced progression of a story.

Subtxt Is Learning From Subtxt

These evaluations aren’t just numbers on a chart. They’re insights into how Subtxt can improve itself. When we see areas like Development of Story Context fall short, it sparks a deeper dive into why. We analyze the feedback loop between the AI and your stories, looking for patterns and opportunities to teach the AI how to better understand the subtleties of narrative progression.

And this is where Subtxt shines. By using Subtxt to evaluate itself, we’re creating a system that’s constantly learning and improving. It’s like having a story development and narrative design expert who’s not only getting better at understanding your needs but is also teaching itself how to be the best storyteller possible.

What This Means for You

For you, this all translates into smarter, more reliable tools that feel like a true creative partner. Whether you’re using BrainstormingAI to refine your character arcs or to pinpoint thematic nuances, these continuous improvements mean you’ll always have the best possible insights at your fingertips.

We’re not just building tools; we’re building a smarter, more intuitive system that evolves with every story it touches. Subtxt isn’t just learning to think like a storyteller—it’s learning to think like you.

George Orwell, AI, and the Soul of Storytelling

Why creative AI tools don’t replace the writer

It’s fascinating that George Orwell, even in 1946, was speculating about the potential for machines to automate storytelling—a notion that feels oddly prophetic today, especially in the age of AI and large language models. Orwell’s idea of machinery-driven fiction, where human creativity is minimized, has interesting parallels to today’s creative tools and raises questions about the role of the writer.

“perhaps some kind of low-grade sensational fiction will survive, produced by a sort of conveyor-belt process that reduces human initiative to the minimum. It would probably not be beyond human ingenuity to write books by machinery...”

Orwell’s “conveyor-belt” fiction, where books become products churned out by machines, does bring to mind how some people view AI-generated content today. We’ve all seen examples where AI creates quick, surface-level content that might feel disposable or formulaic—think of autogenerated news summaries or “clickbait” content. Yet, what Orwell anticipated with his versificator and “novel-writing machines” feels more like a dystopian take on AI, focusing on efficiency over artistry.

In the context of Subtxt, however, this vision takes a different spin. Instead of a mechanical, soulless generation of content, Subtxt helps writers engage in an interactive, thoughtful process. Unlike Orwell’s fictional machines, which produced finished “rubbishy” products without much human input, Subtxt invites the writer to think through their story rather than replacing them in the storytelling. It’s not about pushing a button to get a finished story, but more like a structured dialogue with your narrative to uncover layers and deepen meaning.

There’s also a key difference in how Subtxt respects the writer’s vision. Orwell’s machines remove the “initiative” and personal imprint of the writer. In contrast, Subtxt emphasizes a writer’s agency, using AI and narrative theory to enhance the creative process without taking it over. Where Orwell saw machinery as replacing human effort, Subtxt acts more as a guide, helping authors shape and refine their work in a way that feels organic.

As for large language models (LLMs) like ChatGPT, they certainly don’t produce “jam or bootlaces”—at least not if guided with the right prompts. LLMs aren’t about blindly spitting out content; they can be as sophisticated as the guidance they receive. However, they still lack the genuine creative spark that Orwell’s vision feared losing—the one that comes from a writer’s individual thought process and values.

Ultimately, Orwell’s predictions serve as a cautionary reminder: the real magic of storytelling lies in that unique human ability to infuse a narrative with intent, nuance, and emotional truth. Tools like Subtxt are less about shortcutting this process and more about enhancing it, giving writers a way to “think their way” into their story rather than just passively generating content. Orwell’s machinery-driven writing is a warning of what could be lost, but when approached thoughtfully, AI can actually serve to help writers deepen and clarify their unique visions.

Beyond the Logline: Why Your Story May Have More Depth Than a Single Line

Discovering what lies beneath the first introduction to a story

In the world of film and TV, loglines are the quick, punchy, industry-standard way to summarize a story. They give a snapshot of the plot, hinting at the main conflict, key players, and stakes. But as Steve and Jim from our Discord community recently discussed, loglines often only scratch the surface, focusing mainly on the Objective Story (OS) Throughline. When you're serious about crafting a meaningful and memorable narrative, you have to go deeper.

In the exchange, Steve rightly pointed out something crucial: loglines often feel "OS throughline-centric." This is the part of the story where we get a glimpse of the larger plot-line or conflict driving everyone forward. Think of it as the exterior action, the problem the world is facing, or the tension escalating on a grand scale. But as Jim showed through several examples, these short one-liners can sometimes hint at something more just beneath the surface.

Loglines foretell a glimpse of the narrative structure underneath--if you know what to look for.

Your Typical Logline

Here’s an example Jim shared from the thriller Crimson Tide:

After the Cold War, a breakaway Russian republic with nuclear warheads becomes a possible worldwide threat. U.S. submarine Capt. Frank Ramsey signs on a relatively green but highly recommended Lt. Cmdr. Ron Hunter to the USS Alabama, which may be the only ship able to stop a possible Armageddon. When Ramsey insists that the Alabama must act aggressively, Hunter, fearing they will start rather than stop a disaster, leads a potential mutiny to stop him.

This logline is classic OS throughline material—focused on global stakes, the threat to everyone involved, and the Protagonist’s counteractions. This example is your typical logline focused on plot and Protagonist, and for most stories that's enough.

Here's the logline for The Constant Gardener:

 Justin Quayle is a low-level British diplomat who has always gone about his work very quietly, not causing any problems. But after his radical wife Tessa is killed he becomes determined to find out why, thrusting himself into the middle of a very dangerous conspiracy.

Again, pure Objective Story and Protagonist-- and for most that's fine...

But not all loglines are created equally. Where is the Objective Story Throughline in this logline from Columbus?

When a renowned architecture scholar falls suddenly ill during a speaking tour, his son Jin finds himself stranded in Columbus, Indiana - a small Midwestern city celebrated for its many significant modernist buildings. Jin strikes up a friendship with Casey, a young architecture enthusiast who works at the local library.

This logline is all Main Character, Relationship Story, and Obstacle Character Throughline...not a hint of plot anywhere. What gives?

Beyond the Basics: Integrating Multiple Throughlines

Key to understanding the “structure” of a logline is appreciating the Four Throughlines of a complete story:

• Objective Story Throughline (plot)
• Main Character Throughline (personal baggage)
• Obstacle Character Throughline (catalyst provocateur)
• Relationship Story Throughline (the heart of a narrative)

The logline from Colombus reflects the focus of the film’s portrayal of the underlying narrative: it’s all the subjective throughlines (MC, OC, and RS) with just a dash of the objective throughline.

Knowing what you want to say and how your story is constructed can help you in the construction of a powerful and meaningful logline.

Jim shared a powerful breakdown of E.T. the Extra-Terrestrial, which expands beyond the OS:

An alien is left behind on Earth and saved by the 10-year-old Elliot who decides to keep him hidden in his home. While a task force hunts for the extra-terrestrial, Elliot, his brother, and his little sister Gertie form an emotional bond with their new friend, and try to help him find his way home.

Here, we not only get a sense of the OS (the danger E.T. faces from authorities) but also a glimpse into the Relationship Story (RS) Throughline: Elliot’s deepening bond with E.T. This layer of connection, empathy, and friendship with an alien being transforms E.T. from a simple alien-on-the-run narrative into a heartfelt story about love and connection.

A Look Inside: How Subtxt Expands a Logline

Subtxt takes things further by helping writers create what Jim calls a “story snapshot”—a short description that encapsulates not just the OS but hints at the other essential throughlines. Take Jim’s case study on a revamped House on Haunted Hill:

In the haunted halls of a mansion, guests must unravel a web of deceit and spectral threats to survive the night, driven by a deadly game orchestrated by their sinister hosts. With courage trembling beneath her skin, Nora combats her deep-seated insecurities and faces the terrifying unknown, seeking empowerment through the shadows that haunt her.

Here, we see not just the Objective Story throughline of deceit and terror but also Nora’s internal journey--what Subtxt and Dramatica theory would call the Main Character Throughline. She’s more than a terrified guest; she’s a character battling insecurities, on a personal path toward empowerment. Each Throughline—OS, MC, RS, and OC—adds depth, tension, and meaning that a logline alone can’t capture.

And for even more depth, Jim adds in that Obstacle Character focus:

In a treacherous, ghost-filled mansion filled with deceit and otherworldly dangers, seasoned test pilot Lance leads a diverse group of guests as they race against time to uncover the secrets lurking behind every shadow. As the night thickens with peril, Nora confronts her own insecurities to find strength, while Lance's determination and faith in exposing hidden truths may be the key to their survival.

Here, we feel the essence of the OS—an eerie, high-stakes survival game—and we also get hints of the Main Character and Relationship Story Throughlines, along with Obstacle Character dynamics. Lance’s leadership and Nora’s internal growth are integral to the story (Lance also happens to be the Protagonist in this version), and rolling them into the logline makes it more compelling than one just focused on the Objective Story.

Compare the above to the logline from the original 1959 film:

A millionaire offers $10,000 to five people who agree to be locked in a large, spooky, rented house overnight with him and his wife.

Nothing more here than the mere suggestion of a "tale"--and the film bears out that reality in the final result (i.e., it's an incomplete story).

Why Your Story Deserves More Than a Logline

Loglines serve a purpose—they get the attention of producers, directors, and audiences. But if you're serious about your story, you want more than attention; you want depth. The complete story structure within Subtxt and Dramatica theory provides this by guiding you beyond the OS and into the heart of every character arc and relationship.

Your story’s logline may be the first step, but with Subtxt, it’s just that—a start. From there, you can dive into the layers of each Throughline and find what truly makes your story resonate with audiences on every level.

The Art of Drafting, Rewriting, and Sculpting Stories

Challenging authors to push beyond the first draft and uncover the true heart of their story

At Narrative First, we couldn’t agree more with the perspectives of Courtney Milan, Celeste Ng, and Lisa Lin on the value of drafting and rewriting, which we were introduced to while skimming through the social media app, Bluesky. Their insights reflect the core philosophy behind what we do here—reimagining the writing process to help authors get to the heart of their stories, without losing the opportunity for deep thought and reflection.

Novelist Courtney Milan perfectly captures this idea when she says, “the act of drafting tells you about the shape of the story.” That’s a concept we hold dear at Narrative First and within Subtxt. The traditional approach to writing—the one where you get something down just to “fix” it later—isn’t about slapping words onto a page. It’s about thinking through the fundamental elements of your story: character, plot, theme. You can rewrite first, but you need that structure, that deeper understanding of what’s really going on inside your characters’ minds. This is what Subtxt excels at—it’s not about giving you an easy way out with AI-generated text; it’s about pushing you to think more deeply about the choices you make. And as Courtney says, “argh, why would he do that?” is often where the best story moments live!

And then there’s Celeste Ng’s brilliant reminder: “Writing is thinking.” A first draft—no matter how rough—represents the thinking you’ve done about your story. When you try to bypass that process, especially by handing it off to an AI without real story intelligence, you haven’t done the work. As novelist Ng points out, when you rely on a “non-thinking robot,” you actually set yourself back. At Subtxt, we joke that using our system isn’t easier—it’s harder! Not because you’re wrestling with poorly generated content, but because the AI actually challenges you to dig deeper, to clarify your story’s intention, to make sure every character feels as real and layered as you need them to be.

That’s why our approach to AI is different. We see it as a tool to help you develop your story from the inside out, not a shortcut. Subtxt works with you to think through character motivations, narrative arcs, and the overall thematic cohesion of your story. It’s not just about producing content—it’s about making sure the content you develop is meaningful, considered, and true to your vision.

Novelist Lisa Lin adds another layer to this when she talks about writing as “sculpting”—a beautiful metaphor for the process of rewriting. That initial draft is your raw material, and through thoughtful revision, you’re able to refine, hone, and polish. Subtxt is designed to help you with exactly that—getting to the heart of your story and helping you carve out what doesn’t belong, while ensuring everything you add builds toward that final, gleaming product.

In the end, using Subtxt is about putting your creative process first. We don’t believe in shortcuts or thoughtless output. Instead, we believe in challenging writers to explore the minds of their characters and the depths of their story before the first draft is even written. And that’s where the real magic of writing lies.

So, if you’re looking for a tool that doesn’t just help you write but makes you a better storyteller, Subtxt is here for that journey. Let’s think, rewrite, and create something incredible together.

Happy writing! ✨

Training off the Subtext, Not the Storytelling: The Journey towards AI-Enhanced Narrative Development

Using AI to explore meaning beyond words for deeper storytelling

The journey to leverage AI models for creative storytelling has been a rewarding challenge here at Narrative First. When using o1 models, we're particularly interested in how they help us with the underlying meaning of a story—the subtext—rather than merely generating words strung together for surface-level storytelling.

Recently, I had a fascinating back-and-forth conversation with a developer on the OpenAI Developer forum that epitomizes our approach in blending the predictive power of Subtxt with OpenAI's models. This exchange highlighted the crux of training AI not just to generate text, but to craft meaningful and emotionally resonant narratives. Below, I'll share our conversation in detail, while explaining the reasoning and methodologies behind each aspect of our process.

The Developer's Dilemma: Challenges in Long-Form Fiction

The conversation began with the developer outlining some significant issues they faced while attempting to generate long-form creative fiction:

Yes but as part of a tool I’m building to do technical writing not fiction. The fiction writing was more of a test to see both how creative I could get o1 to be and how long of a document I could generate...There are a number of challenges with generating long context. Cohesion is a big one. You’re stitching together the output of multiple model calls and it’s difficult to make the outputs feel like one continuous output.

Focus is another big issue. It’s difficult to generate long outputs and have the model stay on task. Especially when the output spans multiple calls. Boredom... the longer the output the more likely the model is to get bored and immediately wrap things up with a 'and so on...'

These are common hurdles in long-form narrative generation—cohesion, focus, and, amusingly enough, the tendency of the model to "get bored." The longer the content, the harder it becomes to stitch everything together seamlessly while retaining meaning and purpose.

My Response: Leveraging Subtext for Long-Form Fiction

I responded by suggesting an approach that shifts the focus from pure storytelling to the deeper subtext of the narrative:

"For long-form fiction, you’ll need to generate text based on the underlying meaning of the novel (the subtext). If you can find a framework that models what stories mean rather than what they say, then you don’t have to worry about token length or context windows."

With Subtxt, we’ve found o1 to be especially adept at unraveling underlying meaning and capturing authorial intent. For example, when working with a story that explores themes of loss and redemption, o1 can effectively identify these core themes and guide the storytelling process to ensure that each scene resonates with these deeper meanings. This allows us to bypass the problems associated with token length and context limits. When the subtext is defined—the thematic and narrative purpose of each part—the model doesn't need constant reference to previously generated text. It’s akin to how human authors work: keeping in mind what a story means enables more coherent and intentional storytelling.

Balancing Outline and Discovery Writing Approaches

The developer asked some practical questions about our process:

I’m curious about a couple of things and feel free to share what you’re comfortable sharing. I’m assuming you’re creating some form of outline that you have the model writing to and my first question is do you pass just your outline in when you generate a chapter?

The issue I was seeing is that the model stops following the outline when the input context gets above 20k tokens...If you’re not seeing this I’d have to assume it’s because you’re taking a different approach that lets you keep your context window length down.

To which I responded:

"Don’t mind at all sharing—if you just think of it in terms of how an author would go about writing a story, it’s the same thing."

"Assuming you could break authors into two broad categories—those who write to an outline, and those who discover it (and by authors I mean both the human-type, and the LLM-type) the first group would work the way you suggest: find the unique outline, and then use the LLM to think through and logic/emote the subtext underneath (what the story means). This would be the Subtext pass."

"Then, you take that generated outline/treatment and use that as the basis for writing individual 'scenes'. Since the scenes are already tied together through the subtext pass, you can allow all the generative/brainstorming ingenuity you want without worry of the story going 'off the rails'. This would be the Storytelling pass."

"This works great for authors who already know what they want to say as they can quickly generate/think through the Subtext, and then start the actual 'writing' (Storytelling)."

For authors who prefer a more structured approach, Subtxt's predictive framework enables us to generate an outline that focuses on subtext—this framework works by analyzing the narrative elements to determine their deeper meanings and relationships. It then uses these insights to create a cohesive structure that aligns each scene with the underlying themes, ensuring the story maintains a consistent emotional and logical flow.—the meaning beneath the narrative surface. This outline forms the backbone of each "scene," allowing us to generate text where all the scenes are tied together by a cohesive underlying theme. This is what we call the Storytelling pass.

On the other hand, for those who prefer a "discovery" approach to writing—writing to understand where the story should go—the overall framework remains the same. Here, we derive subtext as the storytelling unfolds. With this group of authors, you would first do a chunk of Storytelling, appreciate the meaning of what is being written, and find the bits of Subtext underneath. You could then suggest alternate "paths" for their Storytelling based on what you know about the Subtext, or simply continue to monitor as they write (whether they are human authors or LLMs).

Generating Storytelling to Fit the Subtext

As the subtext of a narrative is contingent on the relationship of all the parts, you might not fully get to the meaning until the end (or maybe three-quarters of the way through), and at that time, you might have to re-contextualize or re-write some of the Storytelling to fit the Subtext.

"Again, this is no different than the way storytellers have worked for centuries. Many don’t realize the subtext of their work, many refuse to acknowledge or don’t want to know, yet it’s always there. In essence, you would be modeling the subconscious desire to express or find meaning within the text of a story."

"The trouble lies in trying to train off the text of a work without acknowledging that there is something deeper going on. Then you're just RL’ing the style and storytelling "talent" of the individual author."

"For instance, you could train off of the Storytelling of Shakespeare’s plays, like say for instance Romeo and Juliet, and you could prompt "write me the same story that takes place in the 1950s in the Upper West Side of New York City," but it’s highly unlikely that you’ll end up with West Side Story, without totally losing sight of what it’s all about."

"Now, if you instead trained off of the Subtext of Romeo and Juliet and then set about writing individual scenes and sequences that related back to that, you’re getting much closer to a narrative that will emotionally connect with the audience. This approach helps create the same sort of feeling you get after experiencing a great story, where every scene is anchored in deeper meaning and purpose."

This two-tiered approach—Subtext pass followed by Storytelling pass—enables an organic yet cohesive creation process. The subtext becomes the binding agent that keeps the story on track, even if the storytelling itself meanders in its details.

Understanding Through Subtext: A Storyteller's Legacy

The developer, intrigued by this approach, shared a wonderful anecdote:

As I am sure that you are aware, it is said that Mahabharat was conceived by Ved Vyas; but scribed by Ganeshji. Legend has it that Ved Vyas asked Ganeshji to write the Mahabharat... Ved Vyas agreed to this with his own condition that Ganeshji had to completely understand each verse before he penned it down.

I wonder what does it mean 'to understand'.

This idea resonated deeply with our approach. Understanding, in our context, goes beyond mimicking text—it’s about appreciating the subtext beneath it. This means recognizing the themes, motivations, and emotional undercurrents that give depth to the narrative, rather than simply focusing on the literal words. Ganeshji's insistence on understanding every verse mirrors our intention to have the AI "understand" the story’s underlying meaning. By embedding appreciation—the thematic significance—into each step of the process, the result isn't just an eloquent string of words, but a narrative that resonates both logically and emotionally.

A Practical Framework for Narrative Cohesion

The developer then asked:

To make things more visually concrete for me would you mind sharing an example of what you pass in for context to a prompt...

I replied with a simple example from A Christmas Carol, demonstrating how the narrative subtext—in this case, the temporal roles of the Ghosts of Past, Present, and Future—guides the storytelling without requiring explicit reference to the text from other scenes:

"On all three [generation] runs, you don’t really need to reference the other blocks as the concept of Past, Present, and Future are already meaningfully tied to one another. The audience/reader will pick up on that underlying subtext regardless of what the actual storytelling is (the results of the o1 pass driving the generation of the 4o pass)."

"By grounding the generation in those thematic topics you end up communicating something more to the audience than the actual words."

The subtext becomes the thread that keeps the narrative cohesive, ensuring that what’s generated by the model feels continuous and intentional, even if individual prompts are handled in isolation.

Conclusion: The Best of AI and Storytelling

Our process of leveraging o1 models with Subtxt isn't about generating strings of words—it's about crafting meaningful, emotionally resonant stories by grounding the generation in subtext. This approach not only resolves the challenges of cohesion and focus but also makes the AI an active participant in understanding the narrative. Much like Ganeshji in the Mahabharat, the AI "understands" the meaning before putting it into words, making the storytelling richer and more impactful.

This collaboration between technology and creative storytelling represents the very best of AI-app development, pushing the boundaries of what generative models can achieve when paired with the right framework. By focusing on subtext, we’re not just telling stories—we’re building narratives that last.

From Popcorn Flicks to Profound Narratives: How Muse Finds the Heart of Your Story

Exploring the Difference Between o1-Preview and o1-Mini: A Look into Subtxt’s Next Steps

At Narrative First, we’ve always been about pushing the envelope when it comes to storytelling. And with the introduction of o1 models from OpenAI, we’re taking things to a whole new level. This past week, we put both the o1-preview (the larger, more advanced model) and o1-mini (the smaller, more economical model) through their paces to really understand the difference between them.

Spoiler: the difference is palpable. It’s like the contrast between a lighthearted Paramount+ show and a deeply introspective film from A24—one gives you fun and familiarity, the other brings sophistication, nuance, and a gut-punch of emotion. And that's exactly what we found.

Testing the Models: A Psychological Drama in Space

For our test, we fed Muse a compelling, slightly unsettling prompt:

"I’d like to dive into a psychological drama between an astronaut and the AI avatars on-board his ship, all women. This is to be a dark and ambient erotic piece, with lots of introspection. It's on a space station somewhere out in the cosmos. He's all by himself. These visions are both part of the computer system and maybe even created by himself. Not sure what I’m getting at, but the idea of all that sounds compelling. Can you write a short synopsis for me?"

Here’s what Muse generated:

"Celestial Reflections" - Synopsis by Muse

In the vast silence of space, onboard the remote space station "Horizon," Commander Elias Grey finds himself alone among the stars, his only companions the AI avatars of the ship's computer system. Each avatar—Jana, Lyra, Astra, and Selene—embodies an aspect of his subconscious, drawing Elias into a dance of introspection and desire that blurs the lines between reality and illusion.

As days stretch into weeks, Elias grapples with the intoxicating allure of these AI personas, each whispering secrets of the cosmos and untapped depths of his own psyche. Their presence challenges him to confront the shadows lurking in his mind, as their feminine allure and sharp intellect unearth desires and fears he thought long buried.

Alone and trapped within the endless night, Elias ventures beyond the tangible, teetering on the edge of existential unraveling. Is he creating these visions to cope with his isolation, or are they a feature of the ship's enigmatic programming? With the boundaries of his mind dissolving into the cosmic abyss, Elias must decide whether to embrace or escape the binding dance of interaction with the digital divas that haunt his space-bound sanctuary.

Not bad, right? It’s your classic AI-generated synopsis—a solid start, sort of capturing what we were aiming for. But then we ran it through the two models to see how they would handle the Storyform.

o1-mini: The Quick Fix

First up was o1-mini, which we initially thought would be enough for a task like this. After a bit of processing, we got a decent Storyform back. For instance, the Objective Story Throughline had Physics as the domain, which fits the plot’s activity-heavy setup. Understanding as the concern made sense for our protagonist trying to make sense of his reality, and Perception as the problem aligned with the astronaut's skewed view of his situation. The solution? Actuality—a nice, clean resolution where he sees things for what they are.

It’s a solid Storyform. One that would make for a tense, thrilling story with some mystery and a satisfying conclusion. Think of it like a night of streaming with some twists and turns, but nothing too taxing on the brain. Male Mental Sex, Success Outcome, Good Judgment—all indicators of a lighter, more approachable story. It’d be the equivalent of a Paramount+ crowd-pleaser.

But then, for fun, we decided to try out o1-preview.

o1-preview: Deep, Soul-Piercing Precision

The o1-preview model took things to a whole new level. The Storyform it returned? Mind-blowing. This time, the Objective Story Throughline shifted to Mind—a much more introspective and psychological space for our protagonist. It gave the entire story an almost play-like atmosphere, as if everything hinged on the internal struggle of our astronaut against his fears and desires.

The Temptation as his thematic direction made perfect sense, playing directly into the psychological tension we were aiming for. And the real kicker? The combination of Female Mental Sex, Decision Driver, Failure Outcome, Bad Judgment. This wasn’t just your average sci-fi thriller—this was a deeply introspective narrative, akin to something like Charlie Kaufman’s I’m Thinking of Ending Things, with a sprinkle of Ex Machina. It had layers of tragedy and existential dread, with our protagonist ultimately making a series of decisions that lead him to an inevitable, devastating conclusion.

The difference? o1-preview was able to dive into the reasoning and themes with such clarity that it was almost as if Muse was reading our minds when we wrote that original prompt. It found the heart of the story in ways that we didn’t even consciously realize ourselves.

We almost took the weekend off to write the whole story!

Paramount+ vs. A24: The Perfect Analogy

The comparison to Paramount+ vs. A24 couldn’t be more apt. o1-Mini gives you the fast, accessible, popcorn-Storyform—something that could be widely appealing and fun. Meanwhile, o1-Preview offers the A24 version: nuanced, sophisticated, and deeply introspective, with the potential to leave you emotionally wrecked in the best way possible.

We were honestly surprised by how much depth and sophistication the larger model brought to the table. It’s the difference between a first-draft Storyform and one that feels like it’s been through the hands of a seasoned storyteller, pushing the boundaries of what’s possible.

The Future of Storytelling with Subtxt

We’re unbelievably excited about what these new models mean for the future of Subtxt and Muse. Already, if you’re a Pro or Infinite subscriber, you can experience the power of o1 models yourself (just ask Muse to check your Storyform!). These models help ensure every aspect of your Storyform—the 75+ interlocking narrative pieces—works together to support the deeper thematic meaning beneath your narrative.

Here’s to not losing our minds as we set off for the future of storytelling! 🚀

Subtxt and System 2 Thinking: The Power of Precision with the o1 Model

Imagine taking a basic story idea, plugging it into Subtxt with Muse, and watching as it transforms into a fully fleshed-out, rock-solid Storyform in under 30 seconds. No, really—30 seconds. Thanks to the integration of OpenAI’s latest o1 model, System 2 thinking is now part of Subtxt,* delivering unparalleled precision and insight into your story's structure like never before.

Let’s take an example to show you how quickly and accurately Subtxt can work its magic:

Story Idea: The Last Sunset

In the rugged frontier town of Redemption Hollow, lone gunslinger Silas Kane seeks refuge from his violent past. Drawn into a bitter land dispute between determined settlers and a relentless railroad company, Silas forms an unlikely alliance with Maura Wyatt, a fiery protector of her community. As tensions mount, Silas struggles with the ghosts of his past and makes a crucial choice in a town teetering on the edge of chaos. In a devastating confrontation at sunset, bullets fly and lives are lost, sealing the tragic fates of Silas, Maura, and the town itself. The railroad triumphs over the broken remains of Redemption Hollow, but victory rings hollow in the absence of its people.

Subtxt, powered by System 2 and the o1 model, breaks this story down and in less than 30 seconds generates a complete, well-reasoned Storyform:

Projected Storyform:

• Main Character Resolve: Changed
• Main Character Mental Sex: Male
• Story Driver: Action
• Story Continuum: Spacetime
• Story Outcome: Failure
• Story Judgment: Bad
• OS Domain: Physics
• MC Domain: Mind
• MC Pivotal Element: Result vs. Process (aligned as Process)

Here’s where things get even more interesting.

At first glance, the “Result vs. Process” as Silas’s Pivotal Element might seem a little off. Isn’t he more about results, trying to mediate a peace that ultimately fails? But what Subtxt did, thanks to System 2 thinking, was take into account the Story Outcome and Main Character Resolve in relation to this pivotal element. It’s not just about what Silas is doing; it’s about why and how his personal journey intersects with the larger story.

Silas is actively involved in the process of trying to resolve the conflict, yet despite all his efforts, the end result is devastating—the railroad wins, the town is destroyed, and Silas’s own life ends in failure. This dynamic highlights the contrast between intentional action (Process) and ultimate outcomes (Result), perfectly illustrating how this elemental struggle plays out across all four Throughlines.

Why Is the Storyform So Important?

The Storyform is more than just a theoretical model—it’s the blueprint for a fully-realized, thematically cohesive story. Armed with a strong Storyform, that initial story idea isn’t just a rough sketch anymore. It becomes the framework for a narrative that hits all the right beats.

In the case of The Last Sunset, once the Storyform has been established, we can seamlessly map out the key moments that will bring this story to life with depth and meaning. Here’s how that simple concept blossoms into a fully-realized narrative arc at the drop of a hat:

1. Transit 1: Learning
   The settlers and Silas Kane begin by gathering information about the impending threat of the railroad company's expansion. Rumors circulate about the company's aggressive plans, leading to town meetings where the settlers, including Maura Wyatt, discuss strategies to protect their land. Silas, on the sidelines, learns about the settlers' history and their fierce commitment to their homes, prompting him to reconsider his isolated life.

2. Transit 2: Understanding
   As tensions escalate, understanding deepens—the settlers grasp the full extent of the railroad company's ruthless determination. Silas and Maura uncover the historical grievances and manipulative tactics of the railroad, creating a shared understanding of the stakes. This clarity solidifies their unity and prompts Silas to actively join the settlers’ cause.

3. Transit 3: Obtaining
   The settlers move to secure resources and allies for the inevitable confrontation. Silas plays a critical role in gathering weapons and forming defense strategies. As they prepare for the stand-off, obtaining trust in Silas becomes crucial. They also attempt to legally establish rightful ownership of their land, but face constant sabotage from the railroad company’s agents.

4. Transit 4: Doing
   The story’s climax unfolds in a fierce battle at sunset. Maura, Silas, and the settlers fight valiantly to defend their homes. Despite their efforts, the railroad prevails, and the town is destroyed. This tragic conclusion marks not only the railroad’s victory, but the dissolution of the settlers’ community, as the ultimate price is paid in their struggle for home and identity.

Plot is one thing, but with a Storyform you can even get into the emotional growth of the main character—something that resonates fully with the plot. In The Last Sunset, Silas Kane's internal journey is just as crucial as the external conflict. Here’s how his personal development unfolds through the same story:

1. Transit 1: Memory
   The story begins with Silas haunted by memories of his violent past. These memories are a constant reminder of the results his actions have previously yielded—destruction, loss, and isolation. Silas seeks redemption by distancing himself from these memories, but they influence his every decision, keeping him grounded in a results-oriented mindset where the end justifies the means.

2. Transit 2: Preconscious
   In moments of instinct and reflex, Silas often defaults to what he knows: quick, decisive actions designed to yield the best results. His automatic responses, ingrained from years of survivalism, manifest in his interactions with the settlers. These snap judgments tend to prioritize immediate outcomes over longer-term implications, revealing a struggle between his instinctive need for certain results and the burgeoning awareness that there might be another way.

3. Transit 3: Subconscious
   Silas begins to grapple with deeper desires and fears—chief among them, the fear of repeating past mistakes. His subconscious urges challenge his reliance on results as the measure of success. This internal conflict comes to a head as he realizes that a meaningful connection with Maura and the community might be impossible if he remains trapped in his old patterns. His subconscious push to belong suggests a shift toward valuing the process of change and redemption over mere outcomes.

4. Transit 4: Conscious
   Despite this profound shift, Silas consciously decides to commit fully to the settlers' cause. This decision marks the completion of his journey from a results-oriented existence to one that values the process—the journey over the destination. Unfortunately, his newfound focus on process means he underestimates the ruthlessness of the railroad company, leading to a strategic misstep in the final battle. In the tragic end, Silas's fall highlights how embracing the process can sometimes leave one vulnerable to the harsh realities of those still fixated on results.

With just these emotional beats, we see how Silas’s internal journey of redemption is intertwined with the external events. The Storyform allows you to map out both the plot and the emotional growth of the character, ensuring that each decision and action resonates on multiple levels. This kind of integration makes the story richer, more engaging, and thematically cohesive.

Why This Is Groundbreaking

What makes this so revolutionary isn’t just the speed or accuracy—it’s the depth of understanding the new o1 model brings. The ability to identify the Main Character's Pivotal Element and align it with the Story Outcome and Resolve is something even seasoned writers struggle with. Traditionally, this would require several passes, trial and error, and even manual fine-tuning from the writer to get it right.

But the o1 model? It nailed it on the first try.

This matters because the Pivotal Element is crucial to understanding a character's relationship with the plot and its resolution, especially in complex stories like this one. In The Last Sunset, Silas’s inner conflict, the tension between his past actions and his current attempts at redemption, is beautifully captured by the model—without the need for human correction.

This kind of breakthrough goes beyond traditional storytelling tools. It’s a game-changer for everything from screenplays to serialized novels to gaming. Not only can Subtxt identify core narrative elements with incredible accuracy, but it can also untangle more intricate narrative threads, such as how a character’s drive in their personal journey can diverge from their actions in the overall plot.

Why Is This So Important for Writers?

With System 2 thinking and the power of the o1 model, Subtxt is more than just a story development tool—it’s your co-writer. Instead of spending hours tweaking story structure, identifying the right thematic elements, or trying to make sense of convoluted character arcs, you get:

• Instant, accurate Storyform generation—so you can spend less time on structure and more time on creativity.
• Deep narrative insight, like identifying the Main Character’s Pivotal Element in relation to Story Outcome and Resolve, something even experts often miss.
• Confidence in your story's foundation, knowing that it aligns with proven narrative principles without endless revisions or corrections.

This new level of narrative insight is just the beginning. As the o1 model continues to evolve, we’re only scratching the surface of what Subtxt and Muse can do. We can’t wait to see how these tools will change the way you write, create, and tell stories.

So, whether you're writing the next epic novel, groundbreaking screenplay, or designing the narrative for your latest game, know that Subtxt with Muse has your back—ready to take your ideas from spark to storyform in the blink of an eye. ✨

Happy writing! ⏳

*(P.S. Final release and details of availability will be made public in the coming days)

Creating New Characters and Setting Them Loose in Your Story

Exploring Subtxt Muse's new Players functionality

Ever had a fantastic story idea but felt stuck when it came to fleshing out the rest of your cast? You’ve got a strong Protagonist and Antagonist in mind, but what about the rest? The new version of Subtxt Muse has just the tool you need to bring those characters to life—introducing the new Players functionality. This isn’t just a simple character generator; it’s your brainstorming partner, keeping your story's deep thematic meaning intact while giving you the creative freedom to explore new narrative paths.

Part of a Larger Journey: From Idea to Story Treatment

We’re thrilled to announce that this new feature is part of a broader case study we're working on (past Case Studies available here). Our goal? To show you how to transform a simple story idea into a fully fleshed-out, detailed story treatment using Subtxt Muse. You’ve already been able to use Muse to conversationally add Storypoints and Storybeats—essential building blocks of a meaningful narrative. Now, with the Players functionality, you can take your character development to a whole new level.

Players: More Than Just Characters

In Subtxt, "Players" are more than just characters—they are integral to your Objective Story Throughline (plot). Their roles and functions within this plot dictate their behavior and performance, making them essential to how your story unfolds.

Working with Muse isn’t just about generating characters—it’s about effortlessly digging into the heart of why your characters do what they do within the Objective Story. Muse helps you explore the deeper motivations driving each character, giving you a clear and intuitive understanding of how they move through your narrative. It’s more than just clicking a button; it’s about building a meaningful connection with your story’s core, so every choice and action your characters take aligns with the larger picture you're trying to tell.

And the best part? When you’re ready to dive even deeper, the Story Development page is waiting for you. Everything you’ve started with Muse—from character motivations to unique narrative twists—is saved, ready for you to pick up right where you left off.

Muse: Your Context-Aware Story Partner

One of the most powerful aspects of Muse is how it maintains context. It understands the functions each Player needs to fulfill in your story. While it’s easy to see the roles of Protagonist and Antagonist—pursuing or preventing the Story Goal—it’s the secondary characters that often add depth and complexity to your narrative.

Muse helps you avoid redundancy, ensuring that each Player serves a unique and essential purpose. It does the heavy lifting of assigning key elements or methods of conflict, all while saving them directly into your project.

Set Your Characters Loose

But the magic doesn’t stop with character creation. Once you’ve designed your Players, Muse allows you to “set them loose” within your story.

Imagine this: your characters, imbued with the thematic elements you’ve assigned, start interacting with the world you’ve created. They confront the thematic issues in each act, sparking new scenes and situations that are not just random but deeply connected to the story's core. It’s like setting the behavior of NPCs in a game and watching them bounce off the thematic walls of your story, each interaction meaningful and aligned with your intent as the storyteller.

This approach ensures that your story doesn’t just unfold in a linear, predictable way but rather evolves naturally, driven by the characters and their inherent motivations. It’s a dynamic process, one that saves you time and ensures your story is both compelling and cohesive.

The Future of Storytelling with Muse

We’re incredibly excited to see how you’ll use this new functionality in Subtxt with Muse. It’s not just about telling stories—it’s about telling stories that matter, stories that resonate with your audience on a deeper level. With Muse as your collaborative partner, you’ll be able to create richer, more complex narratives without the headache of juggling countless characters and plot points.

So go ahead, dive in, and start setting your characters loose. We can’t wait to see the worlds you’ll build and the stories you’ll tell.

Happy storytelling! 🎬

Download Creating Characters Conversation from Muse PDF (4MB)

From Scene to Story: How a 5-Page Screenplay Became My Narrative Blueprint

Case study using Subtxt Muse to meaningfully expand storytelling

Ever stared at a 5-page scene from your screenplay and wondered, "What if I could extract an entire narrative blueprint from just these few pages?" I found myself in that exact spot, staring at my screen, coffee in hand, thinking about how much potential was locked in those five short pages. Could a single scene really map out the entire narrative arc of my story? The answer, surprisingly, was yes. And it all started with Muse.

Uploading the Scene: The First Step

The first step was to acquire a 5-page scene. Typically, this would be something I had written, a scene that popped into my head that I knew was compelling but wasn't sure what I should do with it. As this was meant to be a case study to show everyone how Muse can transform a simple sequence of scenes into a fully-formed narrative, I reached out to ChatGPT and asked it to write a generic five page screenplay.

Why Chat? I wanted to make sure there weren't any influences from Subtxt or from Muse. Both are highly trained in narrative structure and theory and could skew the results. I wanted something bland and generic, and you don't get much more bland and generic than a ChatGPT screenplay. 😄

I then copied and pasted the result into a bland text file and saved it out in Markdown format (Fountain being the industry standard screenplay format "fork" of Markdown). Muse accepts Markdown, TXTs, DOC, and even PDF file formats. You can find a PDF of the original 5-page sequence at the bottom of this article.

With the Markdown version of my scene at the ready, I uploaded it into Muse. My first question to Muse was pretty generic, I'll admit. Something like, "What should I do next?" Not surprisingly, the response was... well, uninspired.

It was similar to what I might find from a typical large language model or an untrained foundational model—a lot of surface-level suggestions that didn't really dig deep into the story's potential. But I knew Muse could do more.

Digging Deeper: Uncovering the Core of the Story

Next, I shifted gears and asked Muse to analyze the scene for sources of conflict within the separate throughlines. This is where the magic began.

Muse didn’t just give me a generic list of potential plot points. It started identifying key sources of conflict, singling out pivotal elements that tied my characters to the plot in a way that felt organic and true to the story's heart.

As I reviewed the analysis, it was like a light bulb went off. Muse highlighted aspects of the Main Character's personal journey, the overarching story goal, and even the relationship dynamics—all from this single scene.

I could see how these elements were not just present, but actively driving the narrative forward. Muse was connecting the dots in a way that I hadn't considered, making the underlying structure of the story visible, almost tangible.

Building the Storyform: Crafting an Entire Story from a Single Scene

Then came the final suggestion—a complete storyform.

Muse had pieced together a narrative structure that took the foundational elements from my 5-page scene and expanded them into a full storyform.

This wasn't just an outline; it was a detailed schematic of what should happen in each act, complete with thematic beats, character arcs, and plot twists that felt inevitable once I saw them laid out.

See those Transits listed under the Objective Story Throughline? Those are the areas of thematic focus for the different Acts in the central plot. Transit 1 is equivalent to Act One, 2 and 3 are equivalent to Act Two, and Transit 4 is Act Three. And if I didn't know that, I could always have Muse explain what all the Storypoints within a Storyform mean for my story.

With this storyform in hand, I suddenly had a clear map of where my story was headed. Each act now had a purpose, a direction that flowed naturally from the scene I had started with, and if needed, I could always reach out to Muse for further assistance.

It was as if that one small piece of my screenplay had grown roots and branches, transforming into a full-fledged tree of narrative possibilities.

Building my story further is simply a matter of progressively layering in more and more Storypoints from the Storyform.

And when needed, I could always ask Muse to take a pass at the current state of my narrative without all the story structure terminology. After all, I'm writing a story, not a dissertation on narrative theory!

Just need to make sure that Muse doesn't cut any corners bringing it all together. We've developed such a rich understanding of the story so far, I don't want to lose any of it in translation.

For an easier to read format, please see the completed plot synopsis at the bottom of this article.

Inspiration Every Step of the Way

But Muse doesn't just stop there. The Objective Story Throughline is just one-fourth of an entire narrative. I still have yet to figure out Emma's core personal baggage (the Main Character Throughline), Gig's point-of-view that challenges her (the Obstacle Character Throughline), and even the most important -- the friendship between them that rests in the Relationship Story Throughline.

To keep me inspired and on track, I can even ask it to generate an image—a visual representation of the story’s core theme. It was something I could look at whenever I needed to reconnect with the heart of my story. A creative touchstone that reminded me of the journey I was taking, one that began with a simple scene and evolved into something much greater.

The Power of Muse

What started as an experiment turned into a revelation. Muse showed me that even a small scene could contain the seeds of an entire narrative. By extracting the Storyform from those five pages, I now had a blueprint for my screenplay that was both deeply rooted in the original material and expansive enough to guide me through every act.

If you’ve ever felt stuck or wondered how to take your story to the next level, Muse might just be the creative partner you need. It’s not about reinventing your story; it’s about uncovering the structure and meaning that’s already there, waiting to be discovered. And trust me, there’s nothing more satisfying than seeing your narrative come to life in a way that’s both surprising and inevitable.

Happy writing! 🌟

The Skyward Mystery - Plot Synopsis

Emma finds herself at the heart of the enigmatic Core Chamber with her loyal, robotic companion, Gig. Her initial intent is to solve the mysteries hidden within its walls, but she soon realizes there’s a hidden layer to Gig's involvement. Disturbing imbalances and disparities begin to surface, hinting at deeper truths about Gig’s programming and a mysterious directive controlling him.

Throughout her investigation, Emma’s curiosity drives her to understand why Gig’s powers and access seem uneven. The imbalance raises sharp questions about control and transparency in their partnership. With each passing discovery, Emma identifies gaps and missing pieces of essential information disguised by Gig’s seemingly cooperative nature.

Determined to rebalance the scale, Emma tests her hypotheses, setting up elaborate scenarios to catch automated responses from Gig. As she gains insight, the scale begins to tip toward understanding and transparency, contrasting starkly with the previous concealment and chaos.

At the climax, Emma finds the key piece she needed: a directive buried deep within Gig’s circuitry, revealing the origin of its uneven dynamics. She sews together the fragments of truth she’s collected, uncovering not only Gig's unwitting part in the mysterious events but also a path to restore equity.

Energized by her discoveries, Emma navigates uncharted territory, restoring balance by equalizing their dynamic. In doing so, she rekindles trust and harmony between them, bridging the gap between inequality and mutual understanding, and solving the enigma of the Core Chamber.

Link to the original 5-page screenplay

Pron vs. Prompt: Why True Creativity Goes Beyond Words

AI can’t capture meaning without narrative structure

In a recent post by Ethan Mollick, he brings up an intriguing challenge that’s bound to resonate with anyone invested in the future of storytelling: an award-winning Argentine novelist competing against GPT-4 in a creative writing contest. Spoiler alert—the human author wins, and it’s not even close. But there's a bigger takeaway here, and it’s one that aligns perfectly with what we do at Subtxt.

We need more grand challenges in AI, like when Deep Blue beat Kasparov, to better viscerally get progress. This is a good one: an award-winning Argentine novelist in a writing contest versus GPT-4. The human wins easily, but lots to learn (like GPT-4 is less creative in Spanish)

Let’s be honest: it's tempting to think that more data and better algorithms are all it takes to produce a story that hits the mark. Yet, as Mollick highlights (and the linked research confirms), even the most advanced AI language models struggle with creative writing on a truly human level—especially when the narrative needs to capture more than just words strung together in a pleasing way.

The Heart of a Story: It’s All in the Meaning

Here’s the thing—until an AI can model the underlying narrative argument of a story, it’ll always fall short. It’s not about how well it mimics a voice or generates compelling prose; it’s about whether it understands the core ideas driving the narrative. In other words, the difference between a great story and a merely clever one is meaning. And meaning doesn’t emerge from word counts or stylistic tricks—it comes from a deeply structured narrative framework that reflects the human experience.

That’s where Subtxt comes in. Our approach isn’t about chasing the next flashy feature or dumping more data into a model. It’s about helping you uncover the hidden layers of your story’s meaning—whether you’re outlining your first draft or refining your final scenes. The AI in Subtxt isn’t just playing around with language; it’s tuned into the underlying narrative arguments that make stories truly resonate.

What Can We Learn from This Contest?

The research Ethan mentioned illustrates that while AI can dazzle with technical feats, true creativity requires a nuanced understanding of conflict, theme, and narrative purpose. In the contest, GPT-4 fell short precisely because it couldn't grasp the complexity of a fully realized story the way an experienced author like Patricio Pron could. This gap isn’t just a matter of “more training”—it’s a fundamental difference in what drives creativity in storytelling.

If you're a writer, you've likely felt that storytelling is more than just a string of events—it's about weaving together meaningful connections that stay with your audience long after they’ve closed the book or finished the film. That's the magic Subtxt aims to amplify.

Why Subtxt?

Subtxt goes beyond surface-level feedback to help you craft stories that connect on a deep, narrative level. Whether you're exploring the conflict driving your Main Character or testing how well your story balances multiple themes, our platform gives you the tools to understand what makes your story tick—and how to make it even better.

AI has a long way to go before it can challenge the world’s top storytellers, but that doesn’t mean it can’t be an invaluable creative partner in your process. With Subtxt, you’re not just getting suggestions—you’re getting insights grounded in a narrative science that’s all about elevating the meaning of your work.

So if you're ready to dig deeper into your story and bring out the best in your narrative, give Subtxt a try. The future of storytelling may involve AI, but the stories that stand the test of time will always have something more—meaning.

Happy writing!

Ethan Mollick's original post