Skip to content

v1.5.0

Choose a tag to compare

View all tags
@watermark-hd watermark-hd released this 11 Sep 00:37
· 2 commits to main since this release
v1.5.0
4d84ead

Adds a third provider so a PowerPC Mac can front a local / self-hosted LLM.

New: CLAUDE_PROVIDER=openai

Point the agent at any OpenAI-compatible chat/completions endpoint - a
llama.cpp / Ollama / LM Studio server on a faster x86 or ARM machine on your LAN,
for example.

export CLAUDE_PROVIDER=openai
export OPENAI_BASE_URL=http://192.168.1.50:8080/v1   # your server
export CLAUDE_MODEL=<model name the server expects>
export OPENAI_API_KEY=<only if your server needs one>
  • Plain http:// LAN endpoints work (no cert needed).
  • /openai switches to it mid-conversation.
  • Tool use (run_shell etc.) depends on the local model supporting OpenAI
    function-calling; plain conversation works with any model.

Everything from v1.4.x still applies. Full detail in CHANGELOG.md.

Install

Download ppc-claude-agent-v1.5.0.zip, copy the folder to the PowerPC Mac,
double-click Install.command, follow the prompts, then run advisor in a new
Terminal window.

Assets 3
Loading