Hello, I'm Vincent
I made Talespinner because I think everyone has a story in them, and the right tools help more of those stories reach the last page.
My fascination with machine learning started when I programmed a computer to play Othello. Andrew Ng's course on Coursera showed me how much further it could go, and what stayed with me most was the idea of using it to make new things.
Talespinner is where that led. It's a writing studio where your ideas meet an AI's patience: a place for writers, seasoned or just starting, to tinker with a story as much as they like, try things, throw them away, and keep the parts that sound like them.
You hold the pen. Talespinner keeps the world, the characters and the plot in its head, so you can keep writing.
Your stories and your feedback are what move it forward. If something is missing, or could be better, I'd love to hear about it.
The models you can write with
Talespinner gives you several frontier language models, each with its own strengths. Pick one per project, or switch for a single scene.
Claude Opus 5.5 Anthropic
The #1 model for prose on the writing leaderboards and the top-rated fiction craft overall. Unmatched voice, subtext, and character work: reach for it on the chapters that matter most.
Kimi K3 Moonshot
#2 on the creative writing leaderboard with a distinctive, less formulaic voice and a 1M-token window. The alternative when frontier models feel too polished or generic.
GPT-6.1 Sol OpenAI
OpenAI's latest GPT-6.1 model for drafting and revision, with a 1M-token context window for your manuscript, story notes, and instructions.
MiMo-V2.6-Pro Xiaomi
Xiaomi's flagship MiMo model, with a 1M-token context window for your manuscript, story notes, and detailed revision instructions.
GLM 5.3 Z.ai
#3 on the creative writing leaderboard and #1 for long-form novels, with the least clichéd prose of the frontier group — at a fraction of the price. Open-weight and lightly moderated.
Gemini 3.7 Flash Google
Top-3 on Chatbot Arena creative writing at a budget price. Fast, structurally disciplined drafting that keeps plots coherent over thousands of words. The ideal default for everyday work.
GLM 5.3 Flash Z.ai
Near-frontier writing quality — within reach of GPT-5.6 Sol on the leaderboards — at a rock-bottom price, with a 1M-token window. The standout budget pick for drafting at scale.
DeepSeek V4 Flash DeepSeek
The cheapest option in the lineup for rapid iteration and bulk generation. Noticeably below GLM 5.3 Flash in prose quality, so best suited to throwaway drafts and experiments.