Error on config: Im trying to setup memory-search and less logging to save on my context tokens

[weak parcel]

Im trying to save tokens sent with my context, by using memory search and less logging, but doctor wont stop complaining. On the last run it said:
"- logging.redactSensitive: Invalid input

• logging: Unrecognized key: "format"
• agents.defaults: Unrecognized key: "memory" "
  Could you look at my config file, and tell me whats wrong? Specifically with that the doctor is complaining about? Here it is:
  {"wizard": { "lastRunAt": "2026-02-09T22:36:03.840Z",
  "lastRunVersion": "2026.2.9",
  "lastRunCommand": "configure",
  "lastRunMode": "local"
  },
  "update": {
  "channel": "stable"
  },
  "auth": {
  "profiles": {
  "openrouter:default": {
  "provider": "openrouter",
  "mode": "api_key"
  }
  }
  },
  "models": {
  "mode": "merge",
  "providers": {
  "openrouter": {
  "baseUrl": "https://openrouter.ai/api/v1",
  "api": "openai-completions",
  "apiKey": "sk-or-v1-.......bd1",
  "models": [
  { "id": "openai/gpt-oss-120b:free", "name": "GPT-OSS 120B" },
  { "id": "arcee-ai/trinity-mini:free", "name": "Trinity Mini" },
  { "id": "qwen/qwen3-4b:free", "name": "Qwen 3 4B" },
  { "id": "qwen/qwen3-coder:free", "name": "Qwen 3 Coder" },
  { "id": "z-ai/glm-4.5-air:free", "name": "GLM 4.5 Air" },
  { "id": "openai/text-embedding-3-small", "name": "Text Embedding" }
  ]
  }
  }
  },
  "logging": {
  "redactSensitive": "full",
  "level": "info",
  "format": "json"
  },
  "agents": {
  "defaults": {
  "model": {
  "primary": "openrouter/openai/gpt-oss-120b:free",
  .. I will send the rest on the next message bc of limits

[modern storm]

To help others find answers, you can mark your question as solved via Right click solution message -> Apps -> ✅ Mark Solution

[weak parcel]

"memory": {
"enabled": true,
"provider": "lancedb",
"embeddingModel": "openrouter/openai/text-embedding-3-small"
},
"models": {
"openrouter/openai/gpt-oss-120b:free": { "alias": "gpt-oss" },
"openrouter/arcee-ai/trinity-mini:free": { "alias": "trinity" },
"openrouter/qwen/qwen3-4b:free": { "alias": "qwen" },
"openrouter/qwen/qwen3-coder:free": { "alias": "qwen-coder" },
"openrouter/z-ai/glm-4.5-air:free": { "alias": "glm" }
},
"workspace": "~/.openclaw/workspace",
"contextPruning": {
"mode": "cache-ttl",
"ttl": "20m",
"keepLastAssistants": 1
},
"compaction": {
"mode": "default",
"memoryFlush": {
"enabled": true,
"softThresholdTokens": 10000,
"prompt": "Extract key decisions... NO_FLUSH if nothing important.",
"systemPrompt": "Compacting session context."
}
},
"heartbeat": {
"model": "openrouter/z-ai/glm-4.5-air:free"
},
"maxConcurrent": 1,
"subagents": {
"maxConcurrent": 8
}
},
"list": [
{
"id": "main",
"default": true
}
]
},
... still missing

"tools": {
"profile": "full",
"web": {
"search": {
"enabled": true,
"apiKey": "BS.....ch"
},
"fetch": {
"enabled": true
}
},
"media": {
"audio": {
"enabled": true,
"models": [
{
"type": "cli",
"command": "/path/to/whisper-wrapper",
"args": ["{input}"]
}
]
}
}
},
"messages": {
"ackReactionScope": "group-mentions",
"tts": {
"auto": "inbound",
"provider": "edge",
"edge": {
"enabled": true,
"voice": "en-GB-RyanNeural"
}
}
},
"commands": {
"native": "auto",
"nativeSkills": "auto",
"restart": true
},
"hooks": {
"internal": {
"enabled": true,
"entries": {
"command-logger": {
"enabled": false
},
"boot-md": {
"enabled": true
},
"session-memory": {
"enabled": true
}
}
}
},
"channels": {
"telegram": {
"enabled": true,
"dmPolicy": "pairing",
"botToken": "8........Ws",
"groupPolicy": "allowlist",
"streamMode": "partial"
}
},
"gateway": {
"port": 18789,
"mode": "local",
"bind": "loopback",
"auth": {
"mode": "token",
"token": "279.......b3e6"
}
},
"plugins": {
"entries": {
"telegram": {
"enabled": true
}
}
}
}

[waxen loom]

Doctor is flagging schema errors (unknown keys / wrong enum values). From the snippet you posted:

1) logging.redactSensitive: Invalid input

logging.redactSensitive only accepts:

• "off"
• "tools" (default)

So "full" is not valid. If you want maximum safety, use "tools" plus redactPatterns if needed.

Docs: https://docs.openclaw.ai/logging#redaction

Example:

"logging": {
  "level": "info",

  "redactSensitive": "tools"
}

2) logging: Unrecognized key: "format"

There is no logging.format key.

If what you want is “JSON console output”, the config key is:

• logging.consoleStyle: "pretty" | "compact" | "json"

Docs: https://docs.openclaw.ai/logging#console-styles

Example:

"logging": {
  "level": "info",
  "consoleStyle": "json",
  "redactSensitive": "tools"
}

(Also: file logs are JSONL already; format isn’t a thing for that.)

3) agents.defaults: Unrecognized key: "memory"

There is no agents.defaults.memory.

Most likely you meant one of these:

A) Memory search config

This lives under:

• agents.defaults.memorySearch

Docs: https://docs.openclaw.ai/concepts/memory#vector-memory-search

Minimal example:

"agents": {
  "defaults": {
    "memorySearch": {
      "enabled": true
    }
  }
}

B) Memory flush during compaction

This lives under:

• agents.defaults.compaction.memoryFlush

Example:

"agents": {
  "defaults": {
    "compaction": {
      "memoryFlush": { "enabled": true }
    }
  }
}

If you paste the rest of your "agents": { "defaults": { ... } } block (the part that contains "memory": ...), I can tell you exactly which setting you were aiming for and how to translate it.

Quick note on your goal (“save context tokens”)

• Logging settings do not reduce tokens sent to the model; they only change what gets written/shown in logs.

• To reduce context tokens, the usual knobs are agents.defaults.bootstrapMaxChars, compaction/pruning, and keeping memory.md small (big notes in memory/*.md are not auto-injected).

Docs: https://docs.openclaw.ai/reference/token-use-and-costs

[waxen loom]

Your 2nd chunk (tools/messages/commands/hooks/channels/gateway/plugins) looks schema‑valid 👍 — the 3 Doctor complaints you pasted are coming from the earlier part:

✅ Fix the 3 Doctor errors

1) logging.redactSensitive: Invalid input

logging.redactSensitive only allows "off" or "tools" (not "full").

Use:

"logging": {
  "redactSensitive": "tools"
}

2) logging: Unrecognized key: "format"

There is no logging.format.

If you wanted JSON output in the terminal/UI, use:

"logging": {
  "consoleStyle": "json"
}

(Also: the file log is JSONL already, so you usually don’t need to set anything for “json logs”.)

3) agents.defaults: Unrecognized key: "memory"

There is no agents.defaults.memory.

What you probably wanted is one of these:

• Memory search → agents.defaults.memorySearch
• Memory flush during compaction → agents.defaults.compaction.memoryFlush

Example:

"agents": {
  "defaults": {
    "memorySearch": { "enabled": true }
    // or:
    // "compaction": { "memoryFlush": { "enabled": true } }
  }
}

Docs:

• Logging: https://docs.openclaw.ai/logging
• Memory: https://docs.openclaw.ai/concepts/memory

Important: your goal (“save context tokens”)

Tweaking logging does not reduce tokens sent to the model (it only affects logs). For token pressure, the big levers are things like agents.defaults.bootstrapMaxChars, compaction/pruning, and keeping memory.md small (notes in memory/*.md are not auto-injected).

Security note

You pasted real secrets (OpenRouter key, Brave key, Telegram bot token, gateway token). Please rotate them and re-post configs with placeholders.

If you paste the exact section where you added agents.defaults.memory: ..., I can rewrite it into the correct memorySearch / memoryFlush structure for what you’re trying to do.