What's Next for EchoQuest: Our Vision for the Future
EchoQuest started from a simple premise: an audio-first RPG that doesn't care how well you see or how steady your hands are, and doesn't need a tabletop group before you can play. That premise works today. A Game Master narrates every scene aloud and keeps real game state underneath, while you play by keyboard or by voice, with or without a screen reader. Getting here meant a lot of unglamorous fixes, too. This year alone I chased a browser voice that cut off mid-scene and a microphone blocked on my own pages, along with choices that talked over the narrator. I'm proud of where the game stands, bugs and all.
Still, this is only the beginning of what's possible. What exists today is a foundation, a solid one, but a foundation. What I'd like to build over the next two years or so is much larger, and able to serve a far wider range of players. So here's a frank look at where I'm heading: what's close at hand, and the long-range ambitions I'm shaping everything around. Please read all of it as plans, not promises. Priorities shift as I learn, and I'll update this post as things ship (or change).
Near-Term: Deeper World Customization
The thing I most want to give creators is more control over how their worlds behave. On the list:
- Custom sound cue triggers
- Finer-grained rules for how NPCs behave
- Skill check thresholds you define yourself
- Support for non-standard stat systems
- Custom progression schemes
- Tone overrides for specific regions of a world
A few foundations are already there. A Game Bible can define its own classes and backgrounds, set limits on how stats are allocated (a minimum, a maximum, a total point pool) and add custom stats, and the GM's instructions put the world's rules ahead of any default. The rest is still ahead of me, and I'd rather polish each piece properly than rush it out half-wired.
I also want character persistence that runs deeper, so your character's relationships and reputation, plus the history behind them, carry more clearly from session to session. Some of that already exists. Your character's progress is saved on the server, so you can pick up on any device. On top of that, every turn the GM sees where you stand with each NPC you've met (on a scale from sworn enemy to loyal ally, with a short note on why), the lore you've uncovered and a running summary of the campaign so far. The GM does a good job with short-term memory. Now I want long-term history to feel tangible. My goal: by session fifty, the world should mention your reputation casually, the way a barkeep might, instead of you having to remind the GM who you are.
One item from my original list has already landed: importing an existing tabletop campaign. On the Create a New World page there's an Import from Session Notes option. Paste your old notes, whether transcripts, session summaries or world descriptions, and Claude turns them into a Game Bible. It pre-fills the wizard, and you review every field before saving. I built it for tabletop players who'd love to keep a retired campaign going on their own. Got a dusty campaign binder somewhere? This one's for you.
Near-Term: Collaborative Play
Solo play is EchoQuest's foundation, and it'll stay that way. Even so, the feature I'm most eager to build is adventuring with a friend: two players sharing one AI GM and one story. Building multiplayer that works for accessibility is genuinely tricky. Two players might have very different audio setups and screen reader speeds, and their connections won't match either. Work on it has started.
The plan for the first multiplayer release is two-player sessions with synchronised narration, turn-based actions and shared game state, with sessions for three or more players planned for the release after that. I'm paying particular attention to mixed-ability play. A sighted friend and a blind friend playing together should both have a great time, with neither one at a disadvantage. Too many multiplayer games quietly assume everyone uses the same input. The community-run Game Accessibility Guidelines push back on that, recommending that multiplayer games support text chat as well as voice. My version of that idea: one friend can type while the other speaks, and the GM listens to both equally. EchoQuest's multiplayer won't assume a single way of playing.
Medium-Term: Voice-First Interface
Right now, EchoQuest is text-first with audio on top. You read or listen to the narration, then type or speak your action. You can already say voice commands like "pause", "where am I" or "save game", and headset buttons control the narrator. The next step is fully voice-first: you speak, the GM answers in voice, and the keyboard and screen become secondary instead of primary. For accessibility, that would be EchoQuest's biggest leap yet, a fully conversational RPG where the screen is entirely optional.
It's harder than it sounds, because the round trip (your voice in, the AI thinking, a voice back out) has to feel natural. Human conversation sets a brutal bar. A study of ten languages, from indigenous communities to major world languages, found that speakers avoid overlapping and minimise silence, with most turn transitions landing between 0 and 200 milliseconds. No AI round trip gets near that today. Still, I've been chipping away at the delay. Narration starts speaking while the GM's reply is still being written, and the next premium-voice clip is fetched while the current one plays. When voice-first ships, I expect it to open the game to even more people. Players with severe motor disabilities come to mind first, and so do people who'd like to play during a commute or the washing-up. Children and older players for whom typing is a barrier should benefit as well.
Medium-Term: Adaptive Storytelling
I want EchoQuest to learn from how you play over time. If you consistently light up during political intrigue, the GM should start weaving in more of it. If you love quiet character moments, your campaigns should hold more of those. And if you reliably skip past combat, the GM should build stories whose stakes don't hinge on fighting. Personalisation at the level of story is a hard problem. Most game personalisation is shallow ("you killed orcs, so here are more orcs"), yet it's the right problem to tackle.
The opt-in version is simple enough: tell the GM what you enjoy, and it honours that. The harder version is implicit. Over many sessions, the system would notice what you actually engage with versus what you skip, and nudge the world gradually toward your taste. I'm treading carefully here, because there's a fine line between a world that feels tailored to you and a world that just agrees with everything you do, and the second one ruins stories. Researchers studying AI assistants have shown how real that risk is. In one paper, both humans and preference models preferred convincingly written sycophantic responses over correct ones a non-negligible fraction of the time. I won't ship a GM that only tells you what you want to hear. Today's GM is already told that NPCs "dodge questions and lie when it suits them", and that hitting 0 HP brings a real setback. The art lies in making the world responsive without letting it lose its spine.
Medium-Term: A Creator Marketplace
At the moment, creators who publish worlds to the library get visibility and analytics, but no revenue. I want to change that by building a marketplace where world creators earn money when players try their campaigns, with a revenue share for the most-played worlds. The mechanics are still undecided, from revenue-share percentages to payout thresholds, and quality control needs careful thought as well. The principle is settled, though. People who write great worlds for EchoQuest should be able to make a living from it. Right now every bit of the economic upside lands with EchoQuest, and I don't think that's sustainable for a thriving creative community.
Long-Term: Native Multiplatform Apps
EchoQuest is a web app today. It works on phones, but it isn't a native experience. Native iOS and Android apps are on the drawing board, and the goal is full offline downloads of campaigns you've already started. The web version will stay first-class, because a browser tab is still the lowest-friction way in. Even so, mobile and tablet players deserve an experience designed around their devices.
Native apps also unlock things the web struggles with. They'd plug more deeply into each platform's accessibility services, since VoiceOver and TalkBack simply work better in native apps. They'd also allow background audio with the screen locked, plus offline narration for when you travel. The web can stretch further than people think, mind you. EchoQuest already uses the Media Session API, which lets a web page answer headset and lock-screen media keys. Play and pause control the narrator, "previous" replays the last scene and "next" skips the rest of it. But mobile browsers have taught me humility, too. In late May the ambient soundtrack was completely silent on mobile until I tracked it down. These native features will land as the apps mature.
Long-Term: An Open Platform for Accessibility-First Games
My biggest ambition isn't really EchoQuest itself. It's showing that audio-first, accessibility-first design produces better games that more people can play. I want to publish the accessibility patterns I use and contribute to open standards. I also want to keep sharing what I learn about prompting, so other developers can build on it. This blog is a start: I've written up how screen readers work with EchoQuest and how the GM's prompt is built. Blind and visually impaired players have been underserved by the games industry for its entire history. That's a solvable problem. EchoQuest is one answer, but the real goal is a much richer ecosystem of accessible games from many different creators.
I'd also love to work with people doing accessibility research, and with universities studying AI-narrated learning. Groups that use narrative play in therapy interest me just as much. The platform can serve more than entertainment, and I'd like to support those uses without losing focus on the core game.
What We Won't Do
I want to be clear about a few directions I'm not taking:
- No selling player data. Your conversations with the GM aren't used to train any AI model, and they're never sold or shared for ad targeting, as the privacy policy states. Anthropic, which runs the model behind the GM, says that by default it won't use inputs or outputs from its commercial products, its API included, to train its models. One honest footnote: the free tier does show ads, and those ad networks set their own cookies. The privacy of what happens in your sessions is non-negotiable
- No PvP combat. EchoQuest is a narrative, cooperative platform. I'm not building competitive modes or ranked ladders, and player characters won't be fighting each other. That's simply not what it's for
- No NFTs, no crypto, no blockchain anything. Worlds and characters belong to the people who make and play them, not to a token economy
- No dark patterns to push subscriptions. The free tier will stay a complete game, with three campaigns, 60 AI turns a day, browser narration, voice commands and full keyboard access. Paid features should be genuine upgrades, never artificial walls built to force a purchase
A Note of Gratitude
To everyone who has spent time in an EchoQuest world: thank you. Every session teaches me something, and every bit of feedback makes the game better for whoever arrives next. Creators who've uploaded a Game Bible have widened what the platform can be for everyone else, too. In a very real sense, each of you is helping decide what accessible storytelling turns into.
To blind and visually impaired players who took a chance on a new platform: thank you for your patience while I got things right, including the stretch when choices talked over the narrator. And to sighted players who gave audio a go, plus every creator who trusted EchoQuest with their imagination, thank you for spreading the word.
So, which of these would you want first? Tell me. I read every message.
See you in the next session.