v0.5.0

0.5.0 (2026-10-07)

Features

• decisions, jev, local model lifecycle and a model host (#14) (a6b5552)

v0.5.0-decide-and-host.7

0.5.0-decide-and-host.7 (2026-10-06)

Bug Fixes

• openai: cap every request with openai.maxTokens (4b15e31)

v0.5.0-decide-and-host.6

0.5.0-decide-and-host.6 (2026-10-06)

Bug Fixes

• openai: keep a required property non-null in a strict schema (6468f44)

v0.5.0-decide-and-host.5

0.5.0-decide-and-host.5 (2026-10-06)

Features

• llama-cpp: answer many items with schema-valid JSON over one shared prefix (4dc487e)

v0.5.0-decide-and-host.4

0.5.0-decide-and-host.4 (2026-10-05)

Features

• llama-cpp: keep local models loaded across processes in a model host (fdb83e8)

v0.5.0-decide-and-host.3

0.5.0-decide-and-host.3 (2026-10-05)

Features

• llama-cpp: answer decisions from option-letter probabilities (dcd9d59)

v0.5.0-decide-and-host.2

0.5.0-decide-and-host.2 (2026-10-05)

Bug Fixes

• jev: drop the price table entry (7486e9d)

Features

• jev: add the jev provider, a hosted decision provider for TypeSafe's Jev (cab2bcc)
• llama-cpp: add ensureModel, modelState and fits to prepare a local model without loading it (b7a46ee)

v0.5.0-decide-and-host.1

0.5.0-decide-and-host.1 (2026-10-05)

Features

• decide: add decisions as an optional provider capability (4163e99)

v0.4.0

0.4.0 (2026-10-03)

Features

• llama-cpp: run local inference in a worker and fall back from a crashing GPU backend (#13) (415f34d)

v0.3.2

0.3.2 (2026-10-01)

Bug Fixes

• llama-cpp: size the context to the prompt instead of to free memory (#10) (4a31499)