Story-driven text generation for writers and creators
NovelAI is a subscription-based text generation service focused on longform storytelling, roleplay, and creative writing assistance. It provides persistent memory via Lorebook, adjustable generation controls, and in-app image generation (NovelAI Diffusion) for creators who prioritize privacy and custom narrative continuity. Pricing is subscription-based with tiered credit allotments (starter to higher monthly plans), making it best for individual writers, roleplayers, and indie studios rather than free casual use.
NovelAI is a text-generation tool tailored for writers, roleplayers, and creators who need consistent, long-form narrative output and story memory. It uses proprietary narrative-optimized models and features a Lorebook for persistent world details, plus adjustable sampling controls for output tone. NovelAI also offers in-app image generation (NovelAI Diffusion) and export options that suit serialized fiction workflows. The platform is geared toward paying subscribers with tiered credit plans, so the NovelAI text-generation experience is accessible primarily via monthly subscriptions rather than a fully-featured free tier.
NovelAI is a commercial, subscription-first text-generation platform that launched to serve writers, roleplayers, and worldbuilders looking for persistent narrative continuity. Originating as a community-driven project and iterating rapidly, NovelAI positions itself between consumer chatbots and full developer APIs by offering a web app optimized for crafting multi-chapter fiction and interactive roleplay. Its core value proposition is keeping context and character details across sessions using built-in memory systems while protecting user privacy within its subscription model.
Key features center on story-focused tooling. Lorebook is NovelAI’s persistent memory system where users store characters, locations, and facts that the model references automatically during generation. The editor supports adjustable sampling controls (temperature, top_p, repetition penalty) and template-based prompt blocks so users can tune voice and output length precisely. NovelAI Diffusion is the integrated image-generation module (standard export sizes such as 512×512) that creators use for cover art and scene illustrations in the same workspace. The platform also provides prompt management, “Modules” or styles that bias the model toward genres (fantasy, noir, slice-of-life), and API access for automation or third-party integration on paid plans.
Pricing is tiered and subscription centered: there is no permanently unlimited free tier; NovelAI offers entry-level monthly plans for casual users and higher-credit plans for heavy writers. Typical plans give a monthly allocation of generation credits and image credits, with higher tiers increasing text quota, enabling larger context use, and unlocking more Diffusion image credits or faster render queues. Annual billing discounts are usually available. NovelAI’s exact prices and credit quantities change frequently; prospective buyers should check the official pricing page for current monthly amounts and any promotional trials.
NovelAI is used by solo novelists for drafting multi-chapter manuscripts and by roleplay game masters to maintain consistent NPCs and world rules. For example, a freelance author uses Lorebook to keep character arcs consistent across 100k-word drafts, while a tabletop game master uses templates to generate session summaries and NPC dialogue. Small indie studios and interactive fiction hobbyists also adopt NovelAI for quick prototyping. Compared to a general-purpose API like OpenAI, NovelAI emphasizes storytelling features, persistent lore, and an integrated image tool rather than broad API endpoints for application developers.
Three capabilities that set NovelAI apart from its nearest competitors.
Current tiers and what you get at each price point. Verified against the vendor's pricing page.
| Plan | Price | What you get | Best for |
|---|---|---|---|
| Entry | Approx. $9.99/month | Starter monthly text credits and limited image credits, basic Lorebook slots | Casual writers testing story workflows |
| Pro | Approx. $19.99/month | Higher monthly text quota, more image credits, priority queue | Serial writers and active roleplayers |
| Ultimate | Approx. $39.99/month | Largest text and image quotas, API access, extended context usage | Indie studios and heavy creators |
Copy these into NovelAI as-is. Each targets a different high-value workflow.
Role: You are a professional novelist hired to write the first scene of a commercial novel. Constraints: 650–750 words, present sensory detail, begin with an immediate hook in the first line, single POV (third limited or first), avoid heavy exposition—show character through action and small detail, end with a clear inciting incident or unanswered question. Output format: Plain narrative text only, no metadata. Example opening hook: "Rain slicked the alley like a confession." Tone example: wry, atmospheric, slightly tense. Write the scene now, keeping pacing tight and voice distinctive; preserve room for later worldbuilding in subsequent chapters.
Role: You are an industry-savvy literary marketer who writes submission-ready pitches for agents and small-press editors. Constraints: 1) 250±30 word pitch paragraph (compelling stakes, protagonist, antagonist, hook); 2) one 20–30 word logline; 3) one-sentence comparable titles and reason (agent-ready); avoid clichés, use active verbs. Output format: Plain text with three labeled sections: Logline:, Pitch:, Comparable:. Example Logline: "A banker must fake his death to stop a ledger that erases people's names." Tone: crisp, urgent, market-aware. Produce the pitch optimized for commercial adult fantasy/urban fantasy (specify genre).
Role: You are a veteran tabletop game master crafting ready-to-run NPCs for a 3–4 hour session. Constraints: create five NPCs of varied social strata, each with: Name, 2-word tag (role), 1-sentence motivation, 3-character quirks, 2-line voiced sample dialogue in-character, and one short adventure hook (15–20 words). Keep language playable for spoken performance; prioritize distinct voices and quick conflict hooks. Output format: JSON array of objects: {name, tag, motivation, quirks[], dialogue, hook}. Example entry for format: {"name":"Mara Kend", "tag":"bookbinder", ...}. Produce entries for a coastal smuggling town setting.
Role: You are a meticulous worldbuilder creating a Lorebook entry optimized for NovelAI memory and recall. Constraints: include fields Title, Short Summary (1 line), Full Entry (150–220 words), Key Facts (6 bullets), Tags (3–6 keywords), Related Entries (3 names), Known Biases/Conflicts (1–2 sentences). Output format: YAML-like plain text with labeled fields exactly as above. Example Key Fact bullet: "— Treaty of Wrenfall (1493): trade access restricted." Tone: encyclopedic but hookable for narrative prompts. Populate the entry for a ruined skyport city called 'Aeralis'.
Role: You are an experienced interactive fiction designer producing a prototype branching scene for playtesting. Multi-step instructions: 1) Write a 300–420 word scene setup, introducing location, two NPCs, and three player choice types. 2) Provide exactly three distinct player choices; each choice should lead to a 100–180 word branch outcome. 3) Define variables (2–4) that change in each branch with new values and conditional checks. 4) Output a JSON object: {scene, choices:[{id,label,outcome,variable_changes}], variables:{name:initial}}. Few-shot example: {"variables":{"charm":0}, "choices":[{"id":1,"label":"Flatter officer","outcome":"...","variable_changes":{"charm":+1}}]} Use a tense noir inns setting.
Role: You are a senior story architect hired to produce a chapter-by-chapter arc for an 80,000-word novel. Constraints & structure: 1) Divide into 12 chapters; assign approximate word counts per chapter that total ~80k (give numbers). 2) For each chapter provide: chapter title (5 words max), core beat (one sentence), protagonist objective, obstacle, emotional beat, and one line how it advances the theme. 3) Also supply a short three-act summary (3–4 sentences per act) and arcs for two supporting characters (bulleted). Output format: numbered chapter list followed by acts and supporting arcs. Example chapter line: "3. 'Crossing the Docks' — Objective: steal manifests; Obstacle: guard rotation; Emotional beat: guilt resurfaces; Theme: trust vs. betrayal."
Choose NovelAI over OpenAI if you prioritize built-in story memory (Lorebook) and bundled image generation in a single subscription.