5 Classic D&D Campaigns That Inspired EchoQuest
EchoQuest didn't come out of nowhere. It sits on top of roughly fifty years of tabletop RPG design: the campaigns that showed what collaborative storytelling could pull off, and the missteps that exposed its limits. Along the way, a handful of moments proved a narrative game can leave people properly shaken. When I rewrote the AI Game Master's instructions so it would talk like a human GM, I wasn't designing in a vacuum. I was trying to bottle whatever made the campaigns I love actually work, and then asking which parts of that an AI could carry well.
So here are the five campaigns and games that shaped my thinking most directly. None of them involve AI, and all of them came long before EchoQuest. Even so, each one taught me something about how good narrative gaming feels, and those lessons live inside the GM's behaviour today. If you've played tabletop for years, you'll spot the inheritance. If you're new to RPGs, treat this as a half-decent reading list, since these are works worth knowing on any platform you end up playing.
1. Planescape: Torment (1999)
Strictly speaking, this isn't a tabletop campaign at all. It's a video game, and possibly the strongest argument ever made that story matters more than combat in an RPG. Black Isle's Planescape: Torment is famous for its central question ("What can change the nature of a man?") and for companions with real depth, history and moral knots. The protagonist, the Nameless One, wakes on a slab in a mortuary in a city that sits between planes of reality. He has no memory of who he is and no idea why he can't die. Most of the game is spent working out the answer, and that answer is harder and stranger than anything an RPG had tried before (or, arguably, since).
Its lead designer, Chris Avellone, has been candid about how unusual that was. Looking back in 2008, he said the story and quest structure "ended up becoming the primary focus of design", and he even admitted the combat suffered for it. He recalled testers reacting with "Man, this sure is different than the other RPGs that have come through here," which wasn't meant as a compliment at the time. Funny how that works out, isn't it?
What it taught me: a player character's backstory isn't flavour text. It's the engine of the story. The Nameless One's amnesia isn't a gimmick either; it's the central narrative mechanism, the lens every conversation passes through. And the companions you gather are nothing like stock fantasy archetypes. One is a floating, wisecracking skull. Another is a chaste succubus, and a third is a fallen angel working through something heavy. EchoQuest's habit of treating your backstory as an active story input comes straight from this game. The backstory you write sits in the character information the GM reads on every single turn. So don't write it as a bio. Write it as a knot of unresolved questions and let the world keep tugging at it.
2. The Curse of Strahd (1983 / 2016)
Ravenloft is D&D's original gothic horror setting, and Curse of Strahd is its crown jewel. That covers both the 1983 original module, I6, written by Tracy and Laura Hickman, and its 2016 reimagining for fifth edition. The newer version leaned heavily on the original authors. Chris Perkins of Wizards of the Coast said "Without the Hickmans, this project would've died on the vine," and noted that Tracy had been running the adventure for three decades.
What makes it special is Strahd von Zarovich himself. He's a villain with genuine pathos. His psychology hangs together, and his tragic past makes him sympathetic without excusing a single thing he's done. Yes, he's a vampire. He's also a man who fell for the wrong person, struck a desperate bargain, and has spent centuries paying for it while committing fresh versions of the same mistake.
What it taught me: your antagonist needs to be a person, not a symbol of evil. EchoQuest's GM is told that NPCs speak in their own voice with their own rhythm and grudges, that they hold strong opinions and say them plainly, and that they dodge questions and lie when it suits them. A villain run that way feels far closer to Strahd than to a generic dark lord. When you write a major villain into your own Game Bible, I'd add one more habit. Answer these questions about them before anything else: what does this person want, and why do they want it? Then ask what happened to them that made wanting it feel acceptable. Villains are terrifying when they're coherent.
3. Campaigns of Keith Baker's Eberron
Eberron is the D&D setting Keith Baker created for Wizards of the Coast's Fantasy Setting Search in 2002 (it was published in 2004). That contest was a monster in its own right. Wizards' Bill Slavicsek later wrote that the company was "inundated with more than ten thousand submissions." Baker's entry won, and it's built squarely on moral ambiguity.
The premise goes like this. A hundred-year war has just ended in a peace treaty that nobody really trusts. Good nations did terrible things during the war, and the "heroic" factions have blood on their hands. Nobody's clean, and that includes the dragonmarked houses and the church as well as the nations themselves. The whole setting is full of people trying to live with choices everyone made and nobody wants to own up to. Baker has explained the thinking himself: film noir was a major influence, and in his words, "It shouldn't always be easy to decide who the villain is in a scenario… or if killing the villain will solve a problem."
What it taught me: faction design without clean heroes creates the most interesting political choices. Once players know every faction has some legitimacy and every faction casts a shadow, the question stops being "which side is good?" It becomes "what kind of person am I going to be in a world with no clean side?" EchoQuest doesn't have a separate faction engine. Factions live in your Game Bible, and when you upload one, the parser pulls them out along with the world's key tensions for the GM's briefing. My advice for writing them is pure Eberron. Give every group something legitimate it wants that clashes with what the others want, plus one thing it won't do. Then make sure none of them is presented to the player as the canonical good guys. Baker's writing convinced me that grey is more interesting than black and white, and that a GM's job is to honour the legitimate parts of every faction while still letting players pick a side.
4. Matt Mercer's Critical Role (2015–present)
Critical Role didn't invent anything new mechanically. What it did was show millions of people that watching other people play D&D could be genuinely moving television. Moments like Percy's deal with Orthax, Mollymauk's death, Vex's resurrection and the second campaign's Traveler reveal landed hard with viewers who'd never touched a d20. The scale is hard to overstate: in 2019, fans put $11.4 million into a Kickstarter for an animated Vox Machina project that had asked for $750,000, making it Kickstarter's most-funded film project at the time. The show's success reset expectations for what a session could feel like. People who came to D&D after Critical Role expected long-arc emotional payoffs and character moments that earned real tears, with consequences that kept compounding over hundreds of hours.
What it taught me: the emotional power of RPG storytelling isn't reserved for the people at the table. Good performance and careful pacing can move an audience, too. That shaped how I think about EchoQuest's narration. It should be worth hearing as a story, not merely as a game. The GM is told to narrate the way you'd talk across a kitchen table at midnight, and to vary its sentence length on purpose. What I'm after is the care of a performer rather than the efficiency of a referee. The voices matter as much as the words. I once found that NPC voices weren't switching at all, because characters' lines weren't being woven into the narration properly, so every character spoke in the narrator's voice. Fixing that, and matching NPC voices by gender across the whole voice catalogue, made scenes sound like a cast instead of a lecture. Sessions should sound like a story being told, because at their best, that's exactly what they are.
It also taught me something about long-arc patience. Critical Role's most devastating moments often pay off setups from a hundred hours earlier. I won't pretend an AI GM has perfect recall across forty sessions. It doesn't. What it does have is a running campaign summary and NPC relationships that persist. It also has an instruction I'm rather fond of: plant small details now and pay them off turns later, because a reveal hits harder when the player half saw it coming. That's how a seed from an early session can still bloom much later.
5. Apocalypse World (2010)
Vincent and Meguey Baker's Apocalypse World is a tabletop RPG about surviving after the end of the world, but its real contribution was mechanical. The "Powered by the Apocalypse" system has moves trigger from the fiction rather than from declared actions, and it gives the GM (called the MC) an agenda: make Apocalypse World seem real, make the players' characters' lives not boring, and play to find out what happens. The Bakers openly invited other designers to build games on that engine, and as Shannon Appelcline's history at Designers & Dragons traces, it went on to seed a whole generation of indie design. Its core insight is that the GM doesn't need to plan encounters. The GM needs to create situations and ask provocative questions. After that, the job is responding honestly to whatever the players do.
What it taught me: the GM's role isn't running encounters, it's asking hard questions and building situations, then letting players answer them. EchoQuest's GM is built around that idea: set up situations with real stakes, then respond honestly to the player's choices. I didn't paste Baker's "play to find out what happens" into the prompt word for word. The instruction I did write is a close cousin, though: follow the player's creativity instead of forcing them back onto a path. The GM that comes out of it behaves much more like a curious, present GM than a dungeon master grinding through prepared content. It isn't even allowed to decide whether you succeed; a real roll does that, and the GM narrates whatever comes up.
PbtA's split between "soft moves" (foreshadowing, a warning) and "hard moves" (consequences that land) shows up in EchoQuest's pacing too, though I never borrowed the vocabulary. Foreshadowing happens constantly, in those small planted details and the hooks that end each scene. Hard consequences come less often and hit harder: dropping to 0 HP means a real setback, like being captured, or waking hours later with a price to pay. In Apocalypse World, a player's carelessness is what gives the MC licence for a hard move. Here, it's risky actions: they trigger a real roll, and a run of bad ones is how you end up at 0 HP. That rhythm is what keeps the world feeling alive without flattening the player.
What These Five Have in Common
Look at the list together and a through-line appears: every one of these works trusts its players. Planescape trusts them with a strange, melancholy world. Curse of Strahd trusts them with a villain whose motives matter. Eberron trusts them to choose sides in a world without clean ones. Critical Role trusts them to sit through hours of slow character development. And Apocalypse World trusts them to build the story alongside the GM instead of being led through it.
EchoQuest is built on that same trust. The AI Game Master isn't trying to deliver a pre-written experience. It's trying to stay present and attentive to a player who is co-authoring a story nobody could have planned. That's the inheritance. I didn't invent the philosophy; I translated it into something an AI could carry, and I keep refining it as I learn what works. That rewrite was simply the latest pass, and it won't be the last. If you want the longer story of how tabletop habits made it into the GM, I wrote about that in From Tabletop to AI.
Which campaign shaped the way you play? I'd honestly love to know which one I've missed.
