Models registry (generated)

Policy

key	value
include_cleanup_by_default	False
preferred_coding_provider	llamacpp
preferred_general_provider	ollama
reserve_disk_gb	20
reserve_memory_gb	4
tier_balanced_max_ram_gb	31
tier_large_max_ram_gb	63
tier_minimal_max_ram_gb	15

Models

id	provider	role	tier	default	platforms	min_ram_gb	upstream id	automation	description
ollama-general-minimal	ollama	general	minimal	yes	darwin, linux	8	llama3.2:3b	auto	Small general chat model for constrained hosts
ollama-general-balanced	ollama	general	balanced	yes	darwin, linux	16	qwen2.5:7b	auto	Balanced general chat (Qwen2.5 7B)
ollama-general-large	ollama	general	large	yes	darwin, linux	32	qwen2.5:14b	auto	Larger general chat for 32GB+ hosts
ollama-general-max	ollama	general	max	yes	darwin, linux	64	qwen2.5:32b	auto	High-capacity general chat for 64GB+ hosts
llamacpp-coding-minimal	llamacpp	coding	minimal	yes	darwin, linux	8	Qwen/Qwen2.5-Coder-3B-Instruct-GGUF:Q4_K_M	auto	Small coding GGUF for constrained hosts
llamacpp-coding-balanced	llamacpp	coding	balanced	yes	darwin, linux	16	Qwen/Qwen2.5-Coder-7B-Instruct-GGUF:Q4_K_M	auto	Balanced coding GGUF (Q4_K_M)
llamacpp-coding-large	llamacpp	coding	large	yes	darwin, linux	32	Qwen/Qwen2.5-Coder-14B-Instruct-GGUF:Q4_K_M	auto	Larger coding GGUF for 32GB+ hosts
llamacpp-coding-max	llamacpp	coding	max	yes	darwin, linux	64	Qwen/Qwen2.5-Coder-32B-Instruct-GGUF:Q4_K_M	auto	High-capacity coding GGUF for 64GB+ hosts
drawthings-image-balanced	drawthings	image	balanced	yes	darwin	16	flux_2_klein_4b_q6p.ckpt	auto	FLUX.2 Klein 4B (6-bit) — responsive on 24GB Apple Silicon
drawthings-image-large	drawthings	image	large	yes	darwin	32	flux_2_klein_4b_q8p.ckpt	auto	FLUX.2 Klein 4B higher-quality quant for larger hosts
drawthings-image-max	drawthings	image	max	yes	darwin	64	flux_2_klein_9b_q6p.ckpt	auto	FLUX.2 Klein 9B for max-tier Apple Silicon
drawthings-image-minimal	drawthings	image	minimal	yes	darwin	8	flux_2_klein_4b_q6p.ckpt	auto	FLUX.2 Klein 4B (same as balanced; smallest recommended)
fluidvoice-speech-balanced	fluidvoice	speech	balanced	yes	darwin	8	Parakeet TDT v3	manual	Recommended multilingual speech model (app download)
fluidvoice-speech-minimal	fluidvoice	speech	minimal	yes	darwin	8	Parakeet TDT v3	manual	Recommended speech model for smaller hosts
fluidvoice-speech-large	fluidvoice	speech	large	yes	darwin	8	Parakeet TDT v3	manual	Recommended speech model (same upstream default)
fluidvoice-speech-max	fluidvoice	speech	max	yes	darwin	8	Parakeet TDT v3	manual	Recommended speech model (same upstream default)
fluidvoice-cleanup-balanced	fluidvoice	cleanup	balanced	no	darwin	16	Fluid-1	manual	Optional local dictation cleanup model (~3.5 GB)
fluidvoice-cleanup-large	fluidvoice	cleanup	large	no	darwin	16	Fluid-1	manual	Optional Fluid-1 cleanup model
fluidvoice-cleanup-max	fluidvoice	cleanup	max	no	darwin	16	Fluid-1	manual	Optional Fluid-1 cleanup model