win11本地部署openclaw实操第9集-配置 OpenClaw 连接本地模型参数

1.解决上下文窗口问题

OpenClaw 从模型元数据中读取的上下文修改为32768， Ollama 中设置改为32768。

1.1 手动修改配置文件

找到 OpenClaw 的两个 JSON 配置文件：

~/.openclaw/openclaw.json

~/.openclaw/agents/main/agent/models.json

{ "providers": { "ollama": { "baseUrl": "http://localhost:12346/v1", "apiKey": "ollama-local", "api": "openai-completions", "models": [ { "id": "gpt-oss:20b", "name": "gpt-oss:20b", "reasoning": false, "input": [ "text" ], "cost": { "input": 0, "output": 0, "cacheRead": 0, "cacheWrite": 0 }, "contextWindow": 32768, "maxTokens": 32768, "api": "openai-completions" } ] } } }

1.2 重启ollama

(base) gpu3090@DESKTOP-8IU6393:/mnt/c/Users/Administrator$ sudo killall -9 ollama [sudo] password for gpu3090: [1]+ Killed OLLAMA_HOST=0.0.0.0:12346 OLLAMA_MODELS=/home/gpu3090/.ollama/models ollama serve (base) gpu3090@DESKTOP-8IU6393:/mnt/c/Users/Administrator$ OLLAMA_HOST=0.0.0.0:12346 OLLAMA_MODELS=/home/gpu3090/.ollama/models ollama serve &

1.3 重启openclaw

openclaw gateway restart