Verified reference stack
Consultant — Private Email Drafting — CPU 16 GB
Low-cost CPU-only drafting reference using the compact Qwen3-4B package; latency is expected to be higher than GPU profiles.
Open WebUI
Open tool profile →Ollama
Open runtime profile →Qwen3-4B
Open model profile →Fully local chat when only local model providers are enabled
Apache-2.0
Evidence
Continue the decision
Put this reference stack in context
A reference configuration is most useful when you compare it with alternatives and then apply your own hard constraints.
Freshness rule: Material changes set this stack to REVALIDATION_REQUIRED.