Issue: Gateway Token Mismatch & CLI Config Errors on Mac mini (Docker Installation)
Hi everyone, I'm trying to set up OpenClaw on my Mac mini and hitting a few blockers with the connection between the UI and the Gateway. Here is my context:
System Context:
Hardware: Mac mini (Apple M2 Pro), 16GB RAM.
OS: macOS (Sequoia 15.2 / Build 25C56 - reported as 26.2 in logs).
Installation Method: Docker Compose (using openclaw:local image).
Current Problems:
Token Mismatch / Unauthorized:
The UI (localhost:3000) shows "Health: Offline".
docker logs show constant rejection: reason=unauthorized: gateway token mismatch.
I tried setting a token in ~/.openclaw/openclaw.json, but docker inspect shows the container has a different hardcoded OPENCLAW_GATEWAY_TOKEN env var (c8c12...) which seems to override my file settings.
CLI errors:
Running openclaw doctor says:
"Gateway not running." (even though docker ps shows it UP).
Running openclaw channels list fails with:
Invalid config ... - channels.main: unknown channel id: main.
Setup Goal:
I want a clean, local installation using NVIDIA NIM (Kimi k2.5) as the provider.
Configuration Files:
~/.openclaw/openclaw.json (Current Status):
json
{
"agents": {
"defaults": {
"model": { "primary": "nvidia/moonshotai/kimi-k2.5" },
"workspace": "/home/node/.openclaw/workspace"
}
},
"gateway": {
"mode": "local",
"auth": { "mode": "token", "token": "openclaw-rocks" }, // This is being ignored?
"port": 18789
},
"models": {
"providers": {
"nvidia": {
"baseUrl": "https://integrate.api.nvidia.com/v1",
"api": "openai-completions",
"models": [{ "id": "moonshotai/kimi-k2.5" }]
}
}
}
}
Logs (Doctor Output snippet):
โ Plugins โโโโโโโฎ
โ Loaded: 1 โ
โ Disabled: 29 โ
โโโโโโโโโโโโโโโโโโฏ
โ Gateway โโโโโโโฎ
โ Gateway not โ
โ running. โ
โโโโโโโโโโโโโโโโโโฏ
Request: Does anyone know the correct way to align the Gateway Token so the UI connects? And how do I fix the channels.main error to get the CLI working properly? Thanks!
System Context:Hardware: Mac mini (Apple M2 Pro), 16GB RAM.
OS: macOS (Sequoia 15.2 / Build 25C56 - reported as 26.2 in logs).
Installation Method: Docker Compose (using openclaw:local image).
Current Problems:Token Mismatch / Unauthorized:
The UI (localhost:3000) shows "Health: Offline".
docker logs show constant rejection: reason=unauthorized: gateway token mismatch.
I tried setting a token in ~/.openclaw/openclaw.json, but docker inspect shows the container has a different hardcoded OPENCLAW_GATEWAY_TOKEN env var (c8c12...) which seems to override my file settings.
CLI errors:
Running openclaw doctor says:
"Gateway not running." (even though docker ps shows it UP).
Running openclaw channels list fails with:
Invalid config ... - channels.main: unknown channel id: main.
Setup Goal:
I want a clean, local installation using NVIDIA NIM (Kimi k2.5) as the provider.
Configuration Files:~/.openclaw/openclaw.json (Current Status):
json
{
"agents": {
"defaults": {
"model": { "primary": "nvidia/moonshotai/kimi-k2.5" },
"workspace": "/home/node/.openclaw/workspace"
}
},
"gateway": {
"mode": "local",
"auth": { "mode": "token", "token": "openclaw-rocks" }, // This is being ignored?
"port": 18789
},
"models": {
"providers": {
"nvidia": {
"baseUrl": "https://integrate.api.nvidia.com/v1",
"api": "openai-completions",
"models": [{ "id": "moonshotai/kimi-k2.5" }]
}
}
}
}
Logs (Doctor Output snippet):
โ Plugins โโโโโโโฎ
โ Loaded: 1 โ
โ Disabled: 29 โ
โโโโโโโโโโโโโโโโโโฏ
โ Gateway โโโโโโโฎ
โ Gateway not โ
โ running. โ
โโโโโโโโโโโโโโโโโโฏ
Request: Does anyone know the correct way to align the Gateway Token so the UI connects? And how do I fix the channels.main error to get the CLI working properly? Thanks!
The AI that actually does things. Emails, calendar, home automation - all from your favorite chat app. New shell, ready to help. The lobster way.
134,426Members
Resources
Similar Threads
Was this page helpful?
