Write local-usage docs for Claude Code CLI, Kimi CLI, and OpenCode CLI against the local endpoint #6
Notifications
Due Date
No due date set.
Depends on
Reference: haylan/LLM-Server#6
Reference in New Issue
Block a user
Part of #1
Blocked by: #3, #4, #7
Question
Write documentation (README or docs/ file in this repo) covering: how to point Claude Code CLI at the local llama.cpp Anthropic Messages API shim (ANTHROPIC_BASE_URL + dummy ANTHROPIC_API_KEY, any caveats from #3's findings), how to configure Kimi CLI's config.toml to use the local OpenAI-compatible endpoint, and how to configure OpenCode CLI per #7's findings. Include the verified ports/URLs from #4 and any known reliability caveats for using Qwen3.8-27B as an agentic coding backend, so the user knows what to expect before relying on it as a Claude Code fallback.
Write local-usage docs for Claude Code CLI and Kimi CLI against the local endpointto Write local-usage docs for Claude Code CLI, Kimi CLI, and OpenCode CLI against the local endpointResolved. Wrote
docs/coding-cli-setup.md(linked from README) covering all three CLIs — committed onmain(7d8c932).The doc leads with the tool-calling risk warning rather than burying it, and points to the smoke test in #5 before anyone should trust this for unattended agentic work.