Jonathan Singer aa95b0d465 Run Claude on OpenCode's Anthropic endpoint, and name the plan in the UI ...

Claude models are served on /messages, which the OpenAI SDK can't speak, so
those runs go through LiteLLM's Anthropic route instead. Prompt caching moves
with them, since LiteLLM consumes the injection points the raw SDK rejects.

Zen and Go now show up by name instead of both reading 'OpenCode
subscription', and Zen keeps its cost tracked: it bills prepaid credits per
request, so those runs were never actually free.

2026-08-31 21:07:10 +00:00

..

__init__.py

refactor: reorganize core report and tui modules

2026-04-26 14:28:50 -07:00

agents.py

Fix user message retry lifecycle and TUI sync

2026-08-28 23:46:49 +03:00

execution.py

perf: take heavy imports off the startup path and pre-warm them in the background (#1141)

2026-08-20 20:24:08 -07:00

hooks.py

fix(agents): collapse repeated waits queued inside one model turn (#979)

2026-08-06 00:07:30 +03:00

inputs.py

Run Claude on OpenCode's Anthropic endpoint, and name the plan in the UI

2026-08-31 21:07:10 +00:00

paths.py

Local run viewer: email reports, run history, and the platform suite (#813)

2026-07-21 08:13:03 -07:00

runner.py

Isolate MCP connections per task and surface connection status in the UIs (#1181)

2026-08-27 14:10:44 -07:00

sessions.py

fix(sessions): open a sqlite connection per operation, not per thread (#1022)

2026-08-08 16:20:01 -07:00