NovelAI - The AI Storyteller
Anlatan · Writing · Image
NovelAI is a subscription-based AI storytelling tool and anime image generator that runs its own models for fiction writing, image generation, and text to speech narration.
About NovelAI - The AI Storyteller
What Is NovelAI
NovelAI is a creative studio for people who write fiction and make anime-style art. It runs on models the developer, Anlatan, trains in-house, and it splits into three tools under one account: a story editor with an AI co-writer, an image generator with anime-tuned models, and a text to speech narrator. The pitch is freedom and privacy, aimed at uncensored AI writing and a private image tool. The AI won't refuse dark, mature, or experimental writing, and the service doesn't train on your work, which matters to authors who publish what they make. Under that banner, NovelAI doubles as novel writing software for long projects and as an AI art generator for anime fans.
The story editor is where most people start. You write a passage, add some context on the characters and setting, and the AI continues in your style. It also handles text adventure roleplay, where you steer a scene turn by turn. Image generation works the other way around: you describe a scene or list tags, and NovelAI Diffusion builds the picture, with controls for character consistency, pose, and art style that carry across generations.
The limits are worth knowing up front. NovelAI is a paid tool with no permanent free version, only a one-time free trial, and it has no native mobile app, so you use it in a browser or install it as a web app. Its image and text models are tuned for anime and fiction rather than photorealism or general office writing, and Opus subscribers now hit a monthly usage limit on the newest image model, where older models stay unlimited. The service is also capped at 28,672 tokens of context on its largest plan.
Getting Started
- Create an account on the NovelAI website and verify your email to unlock the free trial.
- Use the trial to write in the story editor or generate a few images, then pick a subscription tier.
- Set up a story with context on the characters, setting, and style, or enter tags for an image and tune the settings.
- Run generations, adjust the prompt or the sampling settings, and keep what works.
- Switch on text to speech to hear narration, then revisit the Lorebook or the image tools as your project grows.
Product Information
A quick look at NovelAI - The AI Storyteller's pricing, supported platforms, and performance.
Best for
The users, tasks, and scenarios where this tool fits best.
Users
- Fiction writers who draft novels and short stories and want an AI co-writer that continues their prose and won't refuse mature or dark scenes.
- Anime and manga creators who need consistent characters and a specific art style across many images rather than one-off pictures.
- Tabletop roleplayers and interactive fiction fans who use the text adventure mode to run scenes turn by turn.
- Authors who care about privacy, since NovelAI says it doesn't train on user stories and keeps generations tied to the account.
Tasks
- Drafting a chapter
- Keeping a character on model
- Writing adult or niche fiction
- Narrating a story
- Running a play session
Scenarios
- Working on a long novel where the AI has to remember a large cast, with the Lorebook storing facts so nothing slips between sessions.
- Making a set of illustrations for a story where every image has to match the same character.
- Hosting a solo writing session late at night where you want an uncensored co-writer and full privacy.
- Testing image ideas while streaming, where the Curated model keeps sensitive output from showing up by accident.
Key features
In-House AI Story Models
NovelAI runs its own language models instead of reselling another company's. The current flagship, Xialong, is a 355-billion-parameter fine-tune of ZhipuAI's GLM-4.6 that the team retrained on its own storytelling data, which is why retries give genuinely different results and names come out more varied. Opus subscribers also get older experimental models like Krake and Llama 3 Erato. Because the models are tuned for fiction, prose continues in your style with less of the flat, generic tone a general chatbot falls into.
Story Editor With Full Control
The editor is built for writing, not chatting. You track characters, setting, and style in the story settings, and the AI picks up on all of it. Sliders and sampling controls let you tune how bold, random, or repetitive the output is, so you can push toward wild ideas or lock down a consistent voice. Long projects use the Lorebook to store facts the AI pulls back in as the story grows, which keeps a large cast and its history straight across sessions.
NovelAI Diffusion Image Models, Up to V5
Image generation uses NovelAI Diffusion, a family of models the team built from scratch. The latest V5 is twice the size of V4.5 and was trained in-house on 268,000 B200 GPU-hours. It reads natural language more sharply, accepts prompts in Japanese and other languages, supports alpha transparency, and handles multiple characters, up to 22 distinct ones in a single prompt. The same generation can also render text and lay out a full comic panel, all in one image.
Character Consistency and Style Control
Getting the same character in every image is the hard part of AI art, and NovelAI attacks it head on. Precise Reference pins a character's look so later images match, Vibe Transfer carries an art style from a reference image, and the multi-character prompts place several people in one scene without them bleeding together. These controls make it practical to illustrate a story rather than produce random one-offs.
Text to Speech Narration
NovelAI includes customizable AI voices that read your text back. Every subscription, even the basic Tablet tier, gets unlimited text to speech. You can hear a passage out loud while you edit, or narrate a finished story from the same account. The voices are part of the same private pipeline as the writing and art tools, so narration stays inside one account.
Text Adventure Mode
Text adventure mode turns the story editor into an interactive session. Players type what they do, and the AI answers in prose, running the scene forward. Opus users can run Xialong in this mode, where their typed input goes in as instructions. It's a different way to use the same models, aimed at roleplay and improvised fiction rather than drafting a finished manuscript.
Pros and cons
Pros
- The in-house writing models, including the 355-billion-parameter Xialong, are tuned for fiction rather than general chat.
- Image generation stays consistent across a set, with Precise Reference, Vibe Transfer, and multi-character prompts.
- The AI won't refuse mature or dark writing, which sets it apart from mainstream assistants.
- Unlimited text to speech comes with every paid tier, even the cheapest.
- Anlatan says it doesn't train on user stories, so private and unpublished work stays private.
Cons
- There's no lasting free version, only a one-time trial of 30 images and 50 text actions, and no native mobile app.
- The newest V5 image model now has a monthly usage limit for Opus, where older models stay unlimited.
- Models favor anime and fiction, so photorealism, technical writing, and general office tasks aren't their strength.
- The largest context window, 28,672 tokens, is locked to the $25 Opus tier.
Frequently asked questions
NovelAI is a subscription studio for AI-assisted fiction and anime art. It combines a story editor with an AI co-writer, an image generator tuned for anime, and a text to speech narrator, all under one account and one set of private, in-house models.
Related content
Explore related tools, skills, and articles for NovelAI - The AI Storyteller.
NovelAI - The AI Storyteller Alternatives
Google EmbeddingGemma 2
Google DeepMind · Writing · Image · CodingGoogle EmbeddingGemma 2 is a 740M-parameter open model that turns text, code, images, audio, and video frames into vectors in one 768-dimension space, running on phones and laptops.
MePersona
MePersona, Inc. · Writing · Productivity · MarketingMePersona is an AI unified inbox for creators. It learns your voice, drafts replies to fan messages from 12 platforms, and sends only what you approve.
Fluence
Yay Software (Yannick Huchard) · Writing · Voice & Language · ProductivityFluence is an Android app that reads PDFs, EPUBs, Word files and web articles aloud, follows each sentence as it speaks, and answers questions about the text on the phone itself.