feat: add local inference provider with llama.cpp backend and HuggingFace model management#6933

Merged

jh-block merged 66 commits into

mainfrom

local-models-candle

Feb 19, 2026

Commits on Feb 3, 2026

Local models
Douwe Osinga
committed

Commits on Feb 4, 2026

Update crates/goose-server/src/routes/local_inference.rs

DOsinga
and
Copilot
authored
fix
Douwe Osinga
committed
fmt
Douwe Osinga
committed
Merge branch 'local-models-candle' of https://github.com/block/goose into local-models-candle
Douwe Osinga
committed

Commits on Feb 5, 2026

WIP: local inference debugging - tokenization fix
Douwe Osinga
committed
Merge origin/main to get candle 0.9
Douwe Osinga
committed

Commits on Feb 6, 2026

Make streaming work. tiny model is kinda broken
Douwe Osinga
committed
Simplify?
Douwe Osinga
committed
Only show the "download models" bubble if none have been downloaded
jh-block
committed
Handle EOS during streaming using all EOS tokens from the template
jh-block
committed

Commits on Feb 9, 2026

Commits on Feb 16, 2026

fix: updating to main and starter guard screen (#7241)
michaelneale
authored

Commits on Feb 18, 2026

Commits on Feb 19, 2026