🌶️ Welcome to Spicy Writer

Gemini 4 Hydrogen, GPT-6.2 Sol, some Lite models, story details

2026-11-18

Gemini 4 Hydrogen and GPT-6.2 Sol now up! Google's budget offering and OpenAI's mid tier - they've both been made available over the past couple days, enjoy! Hydrogen has an Exp provider (longer context and output and lower point cost, enabled with gear icon), GPT-6.2 Sol will likely get one soon too.

Story details. The top-right menu in a story has a new Details option: how many messages and characters are on the path you are reading, a token count you can ask for with one click, the story number (the refund page asks for that now, instead of the URL), and when it was created and last updated. Hidden messages are not counted.

New in Lite: Ling 3.2 Flash and ERNIE 5.5 Flash. 1 point each, 32K context. Unfortunately both suffer from pretty heavy rate limits - didn't know it was this bad when adding them, just keep in mind they will often error out.

Opus 4.6 on Experimental should no longer overthink excessively. Opus 4.6 Experimental has had a huge uptick in overthinking recently - not sure why it changed suddenly, but after a lot of trial and error we determined it was at least in part due to defects in Anthropic's own behavior and found a workaround. Ugh sorry I think I misunderstood what the bug was. Still not sure of the cause but pushed a prompt update and increased the point refund when it happens.

The only writing service with easy uncensored access to the latest best frontier AI models like Claude Sonnet/Opus 5.5, Gemini 4 Argon, ChatGPT (GPT-6 Sol), plus hand-picked top open source LLMs, enhanced by an ever growing and improving suite of writing tools!

Spicy: Expert in crafting engaging narratives with a focus on creative freedom and uncensored storytelling.

What is Spicy Writer?

Hey guys, HORSELOCKSPACEPIRATE here. I started creating "jailbroken" story writing prompts and guides for ChatGPT in 2023. My jailbreaks and GPTs have been used millions of times to great success, enabling users to fully stretch their creativity without worrying about being told no, and also enhancing writing quality and length. Sadly, on ChatGPT, OpenAI fights tooth and nail to thwart us with A/B testing and the censorship roller coaster, and it's a difficult battle for my users. Out here, we serve these models over API, which is so much more stable, with so much more control, allowing us to set up powerful writing tools to make your writing experience even better - I'm sorry it took me so long to start my own service, but here we are!

Who is Spicy Writer for?

Anyone who's been told "no" by other AI writing tools. Whether you're writing mature themes, exploring complex subjects, or just want AI that treats you like an adult without constant safety lectures - this is for you.

While we're NSFW-friendly and don't restrict adult content, many users choose Spicy Writer for completely SFW projects. They love what our prompts do to writing quality - the length, depth, and creative freedom that comes from AI that isn't second-guessing itself or softening every edge!

Why Choose Spicy Writer?

Because other services fall short on keeping up with the best bleeding edge models as they come out, or don't even try to carry good models to begin with. Here, you'll have a passionate, highly experienced jailbreaker and context engineer at your back to consistently bring you the latest and greatest, committed to hand-picking the best models actively used and loved by writers and roleplayers, with prompts written by myself and other elite veterans in my network like u/Spiritual_Spell_9469. I spent years providing prompts for free because I truly want people to write what they want, how they want, and be treated like adults without being lectured.

We also have a vision for the best context management and "memory" of any AI writing system, period. We've recently introduced file upload, powered by intelligent hybrid vector search, to enrich outputs. Lorebook and advanced summary features are soon to come, combining into a seamless writing experience that no other service can even dream of touching.

We strive to offer excellent value, giving 150% the context window of ChatGPT or more! We don't make this claim lightly - frontier models are expensive to serve, and we are truly unique in the space in that we actually provide AND jailbreak the newest state of the art models - some of the few notable competitors in this space (not naming any names) have the gall to charge more for much older models, or have no frontier models at all!