Open models like Llama, Qwen and DeepSeek can be downloaded and run on your own computer or server; the usual deployment tools are Ollama (friendly for personal desktops) and vLLM (high-throughput servers). No API bills, no content leaving your network — the cost is needing enough VRAM and compute.
Enclave lets you point any character at a local model: privacy-sensitive characters run locally, intelligence-demanding ones use big cloud models — mix and match.
Try it yourself in Enclave
Open it in your browser — no credit card, no install.