How to Write D&D Session Notes That Actually Help Next Week

Most D&D session notes are written for the wrong person.

They are written for the version of you who just finished the session, still remembers the joke about the suspicious goat, still knows which NPC was lying, still has the rhythm of the night in your head. That person does not need much help. That person was there.

The person who needs help is future-you.

Future-you is tired. Future-you is prepping on a Thursday night after work. Future-you vaguely remembers that the party promised someone something, but cannot remember whether it was the apothecary, the smuggler, or the widow with the cursed orchard. Future-you knows the rogue picked up a key, but not which lock it was supposed to open. Future-you is about to run the next session with just enough confidence to be dangerous.

That is who your notes are for.

Not the historian.

The GM who has to pick the campaign back up next week.

If you just want the copyable version, there is a full D&D session notes template below. The rest of this post is about why those fields matter, because a template only works when it matches the job your notes are supposed to do.

Your Notes Are Not a Novel

A session recap is not a short story.

It does not need polished prose. It does not need a dramatic opening line. It does not need to capture every exchange at the table. If your notes are becoming a second writing assignment after every game, they are going to fail the first week you are busy.

Good D&D session notes are not about writing more.

They are about capturing what needs to survive.

That distinction matters because the average session contains a lot of noise. Good noise, often. Jokes, table banter, false starts, tactical debates, bits that were hilarious because everyone was exhausted and the dice were weird. Some of that belongs in a campaign journal if your group likes keeping one. Most of it does not belong in the working notes you use to run the game.

The working notes have a job: preserve continuity.

If a detail will not affect a future scene, decision, relationship, mystery, inventory state, or player memory, it probably does not need to be in the GM notes.

That sounds ruthless until you realize how freeing it is. You are not failing to document the campaign. You are choosing what the campaign needs to remember.

The Five Things Every D&D Session Notes Template Needs

A useful session notes template does not need twenty fields.

It needs five.

1. What changed?
2. Who changed?
3. What was learned?
4. What was promised?
5. What is unresolved?

That is the spine of the whole thing.

Everything else is optional.

What changed?

Start with state changes. The campaign was one way before the session and another way after it. Your notes should make that difference obvious.

Examples:

• The party found the moonstone key in the chapel crypt.
• Mara took the cursed dagger and has not identified it.
• The town guard now suspects the party was involved in the warehouse fire.
• The goblin scouts escaped and can warn the Ironroot clan.
• The party spent 300 gp on spell components and healing potions.
• The old bridge collapsed and the northern road is blocked.

These are the facts that prevent continuity drift. If you only write a narrative recap, these details get buried. If you write state changes first, your next prep session starts with the campaign as it actually exists now.

Who changed?

NPCs are where session notes usually start to rot.

Not because GMs do not care about NPCs, but because relationships change in small ways. The blacksmith is annoyed, then suspicious, then quietly helpful, then in real danger because the party used his forge as cover for a bad plan. If those shifts live only in your memory, they will flatten over time.

Track NPC state like this:

• Captain Velra: believes the party lied about the warehouse fire. Wants proof before helping again.
• Old Fen: now owes the party a favor after they returned his ledger.
• Sister Maud: noticed Kaelen recognized the chapel sigil and will ask about it privately.
• Tovin: fled Brightwater after the fight. Likely carrying news to the Ashen Court.

You do not need an essay. You need the current state of the relationship, the relevant motivation, and what the NPC is likely to do next.

That is enough to make them feel alive later.

What was learned?

Players learn things out of order. That is part of the fun.

The GM problem is remembering what they know versus what you know.

Write down discoveries separately from events. The party did not just “visit the chapel.” They learned that the chapel sigil matches the mark on the moonstone key. They learned that the priest was hiding letters from the baron. They learned that undead avoid the western crypt, which means something worse may be there.

Useful notes sound like this:

• The party knows the Ashen Chancellor uses silver-threaded messengers.
• They know the mayor lied about never visiting the chapel.
• They do not know the moonstone key opens the observatory door.
• Mara suspects the dagger is connected to her patron, but has no proof.
• The group incorrectly believes Tovin killed the courier.

That last category matters.

False beliefs are campaign fuel. If the party believes something wrong, and you forget they believe it, you lose one of the best sources of future tension.

What was promised?

Promises are where campaigns become personal.

They are also the first thing everyone forgets.

Players make promises constantly. They agree to return with medicine. They tell an NPC they will investigate the disappearance. They swear they will not reveal a secret. They accept payment in advance. They threaten someone and expect that threat to just hang in the air forever without consequences.

Write them down.

Not because you want to punish the players, but because promises are hooks the table already created. You do not need to invent motivation from scratch when the party already left a trail of obligations behind them.

Track promises like this:

• Party promised Sister Maud they would not tell the town guard about the hidden letters.
• Kaelen promised the ghost he would bring the moonstone key to the observatory.
• Mara told Tovin she would kill him if he lied again.
• The group accepted 200 gp from Old Fen to recover the missing ledger.
• The party told Captain Velra they would return before sundown, then did not.

That last clause is the important part.

A broken promise is not just a missed task. Handled well, it is a future scene waiting for its moment.

Separate Public Recap From GM Memory

Most tables need two kinds of notes.

The first is the public recap: what the players can safely read. This is the version that helps everyone remember what happened last time. It should be clear, short, and written from the party’s perspective.

The second is GM memory: what you need to run the world honestly. This includes secrets, hidden motivations, consequences in motion, incorrect player assumptions, and NPC plans.

Do not mix them casually.

If you write one giant recap with everything in it, you either spoil your players or you stop writing down the dangerous parts. Neither is good.

The public version is also a collaboration tool. It lets the table share memory instead of making the GM carry every detail alone, which is the same reason party collaboration matters outside the notes document too.

A public recap might say:

The party found a moonstone key beneath the ruined chapel, fought off three gravebound soldiers, and returned to Brightwater after dark. Sister Maud warned them not to trust the mayor.

The GM memory might say:

The key opens the observatory door. Sister Maud knows this but is testing whether Kaelen recognizes the sigil. The mayor did not send the gravebound soldiers, but he will be blamed if the party repeats Maud’s warning publicly.

Those are both true.

They do different jobs.

If your notes do not separate player-facing memory from GM-only truth, the campaign gets harder to manage every time the party learns a half-truth.

Write for the Next Session, Not the Last One

A lot of GMs write notes backward.

They ask, “What happened?”

That is useful, but incomplete. The better question is:

“What will I need when I prep next time?”

That changes what you record.

You stop writing:

The party talked to the mayor for a while and did not trust him.

And start writing:

The party does not trust the mayor. They have no proof. The mayor noticed Mara watching his signet ring and will stop wearing it in public.

The second version gives you something to use.

Good notes should create next-session handles. When you read them, you should immediately see what can move.

An NPC can act. A clue can resurface. A promise can come due. A faction can respond. A player suspicion can be rewarded or complicated. A piece of loot can matter.

The recap is not the destination. It is the handoff to the person doing next week’s prep.

D&D Session Notes Template

Here is the template I would actually use.

Not the prettiest one. Not the most comprehensive one. The one least likely to become homework.

# Session [Number]: [Title or Location]
Date played:
In-world date:
Party present:

## Public Recap
- 
- 
- 

## State Changes
- 
- 
- 

## NPC Updates
- **NPC Name:** current attitude, new information, likely next move
- **NPC Name:** current attitude, new information, likely next move

## Clues and Discoveries
- The party learned:
- The party suspects:
- The party is wrong about:
- The party has not discovered:

## Promises, Debts, and Threats
- 
- 
- 

## Loot and Inventory Changes
- 
- 
- 

## Unresolved Threads
- 
- 
- 

## Prep Reminders for Next Session
- 
- 
- 

For most tables, that is enough to make next week meaningfully easier.

If you want more, add it carefully. Every field you add is another field you can fail to fill out. A template that gets used imperfectly every week is better than a perfect template you abandon after two sessions.

The “Three Minutes After Session” Version

Sometimes you will not fill out the full template.

That is fine.

The real failure mode is telling yourself you will remember later.

You will not.

So use the three-minute version before you close your laptop, leave the table, or go to bed:

## Session Debrief
- Biggest thing that changed:
- NPC who needs to react:
- Promise or consequence to remember:
- Loot/item/clue gained:
- First thing to prep next time:

That is the whole emergency version.

Five lines.

This tiny version has saved more campaigns than elaborate note systems, because it catches the fragile stuff before it evaporates. You can expand it tomorrow. Or not. Even if you never touch it again, future-you has five handles to grab.

Track Loot Like It Matters

Loot notes deserve their own section because inventory is one of the places where “we will remember” becomes a lie fastest.

The party found three potions, a strange dagger, 200 gp, a silver idol, and a scroll nobody identified yet. By next week, one player thinks the potions went to the cleric, another thinks they are in the party bag, and the rogue has already written the dagger onto their sheet without the curse you planned.

That is how treasure turns into fog: not all at once, but one fuzzy handoff at a time.

At minimum, write:

• what was found
• who took it
• whether it was identified
• whether it is cursed, secret, or incomplete
• whether it belongs to a character or the party
• whether it should matter later

Example:

## Loot and Inventory Changes
- 200 gp from the chapel crypt, currently in party funds.
- Moonstone key, held by Kaelen. Opens observatory door. Party does not know this.
- Unidentified black dagger, held by Mara. Warm near undead. Actually tied to her patron.
- Two healing potions, one to Orin and one still in party bag.

That is the whole difference between a magic item becoming a future reveal and a magic item becoming a line nobody trusts on a character sheet.

If your table already uses a shared inventory tool, this gets easier. If not, your notes are the source of truth, so make the truth specific.

Track NPCs by Pressure, Not Biography

You do not need a full biography for every NPC.

You need pressure.

What do they want? What are they afraid of? What changed because of the session? What are they likely to do next?

That is enough to make them act consistently.

Bad NPC note:

Mayor Hadric: suspicious, older man, fancy ring, seems corrupt.

Useful NPC note:

Mayor Hadric: knows the party found the moonstone key. Afraid they will connect it to the observatory. Will invite them to dinner to measure what they know. Stops wearing his signet ring in public.

The second note gives you a scene instead of a vibe.

That is the standard. If an NPC note does not help you run the next scene with that NPC, it probably needs less description and more pressure.

Track Player Theories

This is the note category most GMs skip, and it is one of the most valuable.

Players tell you what they care about by what they theorize.

If the table spends ten minutes arguing that the mayor is secretly undead, write that down. Even if it is wrong. Especially if it is wrong.

Player theories are useful because they show you:

• what clues landed
• what clues failed
• what the players are excited about
• what assumptions they are carrying
• what future reveals might satisfy or surprise them

You do not have to change the truth to match every theory. That way lies mush.

But if the players are obsessed with the mayor’s ring and you wrote the ring as a throwaway detail, maybe the ring should matter. Not because the players guessed the plot, but because they told you where the energy is.

A good note might be:

Player theories:
- Group thinks Mayor Hadric is undead. False, but they are very invested.
- Mara thinks the moonstone key belongs to her patron. Partly true.
- Kaelen suspects Sister Maud is protecting someone. True.

That gives future-you a map of the table’s attention, which is usually more valuable than another page of polished recap.

How AI Can Help Without Deciding Canon

This is where an AI assistant can be genuinely useful.

Not by deciding what happened. Not by inventing canon while you are not looking. Not by turning your campaign into a content slurry.

By taking messy raw notes and helping you structure them.

You can paste in a chaotic after-session dump:

chapel fight, 3 grave soldiers, mara got dagger, k suspicious of sigil, mayor lied maybe, maud said dont trust him, key found under altar, old fen ledger still unresolved, they forgot velra deadline

And ask for:

• a public recap
• GM-only state changes
• NPC updates
• unresolved promises
• loot and inventory changes
• prep reminders

That is not the assistant running the campaign. It is closer to cleaning the table after the meal: sorting the plates, finding the missing fork, and making sure tomorrow’s breakfast does not start in a pile of last night’s dishes.

You still review every line. You delete what is wrong. You correct names. You decide what becomes official.

I wrote more about this line in AI Dungeon Master Assistant: What Actually Helps at the Table, but the short version is this: AI belongs in the note process when it makes your memory easier to use. It does not belong there when it quietly starts deciding what your campaign means.

If you are using Scrybe, this is the kind of work I think it should be doing: helping you turn the raw material of play into a campaign record you can use later.

From Notes to Campaign Memory

The best version of session notes does not end with a document.

It becomes campaign memory.

That means your recap updates the NPCs. The loot changes the inventory. The discovered faction gets connected to the right location. The unresolved promise stays visible until it resolves. The next time you ask “what does the party know about Brightwater?” the answer includes what happened last Thursday.

That is why I care about lore, characters, inventory, and import as connected parts of the same problem. A campaign is not one document. It is a living state with people, places, items, secrets, and consequences moving through it.

Session notes are the bridge between what happened at the table and what the campaign remembers.

If that bridge is weak, everything downstream gets harder.

The GM forgets. The players forget. The world stops reacting. Old promises vanish. NPCs reset to their default personalities. Loot becomes spreadsheet noise. Mysteries lose their shape.

If the bridge is strong, the opposite happens.

The world feels like it remembers because you remembered what to carry forward.

The Point Is Not Perfect Notes

You are going to miss things.

You will forget an NPC name. You will write “weird chapel symbol???” and have no idea what you meant two weeks later. You will confidently record that Orin took the potion, only for Orin’s player to swear they gave it to Mara. This is fine. The goal is not archival perfection.

The goal is enough continuity to keep the campaign alive.

A useful note system should lower the bar, not raise it. It should make the smallest responsible version easy enough that you actually do it after a long session.

So start small.

Write the public recap. Write what changed. Write the promises. Write the loot. Write the one thing you need to prep next.

That is already better than most campaign notes.

Your future self does not need a novel.

Your future self needs handles.

Give them something to grab.

---

ScryMarket is built around the idea that campaign notes should become usable campaign memory, not another forgotten document. Scrybe, Lore, and Import help turn scattered notes into something your table can actually use. Start with the Scrybe guide, explore Lore, or bring old notes into the Import tool.