Free
A free trial to feel the prose and find your model.
$0
- Try every model
- 2 concurrent chats
- A couple of scenes a day
- OpenAI-compatible API
Open Weights
Open Future
Always in character.
Open by Design
Powerful Open Models for Roleplay
The best open models for roleplay and long-form fiction — always in character, with the memory to hold a long scene. Plug in SillyTavern, or write right here.
Open Weights
Open Future
Community
First
Built for
Long Context
Our Models
View All →Artemis
Live
The community favorite for character roleplay. Expressive prose, strong character voice — serving now in closed beta.
31B Dense | 64k Context | FP8
Join the Beta → 
Vaelico V1
Coming Soon
Our flagship in-house model. A narrative fine-tune built for character consistency and long-form coherence.
50B Dense | 256k Context | NVFP4
Join Waitlist → 
GLM-Steam 106B
Coming Soon
Maximum coherence, deep memory. Among the first to serve it as a hosted API — built for long-form narrative.
106B MoE | 128k Context | NVFP4
Join Waitlist → 
Artemis
Live
The community favorite for character roleplay. Expressive prose, strong character voice — serving now in closed beta.
31B Dense | 64k Context | FP8
Join the Beta → Vaelico V1
Coming Soon
Our flagship in-house model. A narrative fine-tune built for character consistency and long-form coherence.
50B Dense | 256k Context | NVFP4
Join Waitlist → GLM-Steam 106B
Coming Soon
Maximum coherence, deep memory. Among the first to serve it as a hosted API — built for long-form narrative.
106B MoE | 128k Context | NVFP4
Join Waitlist → Why Vaelico?
Creative Freedom
Models that follow your narrative and stay in character — no corporate guardrails breaking the scene.
Built for Roleplay
Tuned for character consistency and long-form coherence — not trivia or yes/no answers. Deeper characters, better prose.
Open Weights
OpenAI-compatible, no black box. Models you can inspect, run and trust — no corporate leash.
Creative Freedom
Models that follow your narrative and stay in character — no corporate guardrails breaking the scene.
Built for Roleplay
Tuned for character consistency and long-form coherence — not trivia or yes/no answers. Deeper characters, better prose.
Open Weights
OpenAI-compatible, no black box. Models you can inspect, run and trust — no corporate leash.
Plug in your frontend
One OpenAI-compatible endpoint. Paste it into the app you already use — no lock-in, no bespoke SDK, no waiting on us to build a client.
- SillyTavern
- RisuAI
- NovelCrafter
- Janitor AI
OpenAI-compatible
- Base URL
- https://api.vaelico.ai/v1
- Auth
- Bearer vk_…
- Frontends
- SillyTavern · RisuAI · NovelCrafter · Janitor AI
Pricing
No lock-in. Connect the frontend you already use in a minute. Free to try — upgrade when you live in it.
Pro
For daily drivers who live inside long scenes.
$15 /mo
- All models, including Vaelico V1
- 6 concurrent chats
- Priority speed
- Room to roleplay all day
Built on Blackwell.
Tech Stack
- > Inference
- vLLM / SGLang
- > Quantization
- NVFP4
- > Hardware
- NVIDIA B200
- > Distribution
- Hosted API
- > API
- OpenAI-compatible
NVFP4 inference on NVIDIA B200.
B200 NVFP4 inference on NVIDIA B200.