Able to import and export Settings

[AverageRed]

Pretty self-explanatory but I've been planning on switching my pc and i don't want to configure it one by one again. It should include

• Settings
• Profiles
• Custom Instructions
• and everything really

[RyzeNGrind]

+1 Any updates on this, I just switched from Windows to NixOS. Really need declarative config mgmt for vscode extensions especially better roocode + vscodium would be good.

[dsent]

It seems that most settings now live somewhere in the state database, right?

It'd be awesome to have them all in a normal config file, probably somewhere under globalStorage/rooveterinaryinc.roo-cline/settings, right where custom_modes.json and mcp_settings.json already live.

[mrubens]

How far does the import/export button in the last section of settings get you?

[aa956]

+1 for editable settings in text (yaml/json/toml) file.

import/export button

Looks good! But there are "id" fields in the exported file (database keys?).

Say, I've added some models manually and exported the settings:

{
  "providerProfiles": {
    "currentApiConfigName": "qwen3-32b-q4-ctx128k",
    "apiConfigs": {
      "qwen3-30b-a3b-q4-ctx128k": {
        "openAiBaseUrl": "http://localhost:8000/v1",
        "openAiApiKey": "none",
        "openAiR1FormatEnabled": true,
        "openAiModelId": "qwen3-30b-a3b-q4-ctx128k",
        "openAiCustomModelInfo": {
          "maxTokens": -1,
          "contextWindow": 128000,
          "supportsImages": false,
          "supportsPromptCache": false,
          "inputPrice": 0,
          "outputPrice": 0
        },
        "openAiHeaders": {},
        "apiProvider": "openai",
+        "id": "7z6b2zsnv6r"
      },
      "qwen3-32b-q4-ctx32k": {
        "enableReasoningEffort": false,
        "openAiBaseUrl": "http://localhost:8000/v1",
        "openAiApiKey": "none",
        "openAiR1FormatEnabled": true,
        "openAiModelId": "qwen3-32b-q4-ctx32k",
        "openAiCustomModelInfo": {
          "maxTokens": -1,
          "contextWindow": 32768,
          "supportsImages": false,
          "supportsPromptCache": false,
          "inputPrice": 0,
          "outputPrice": 0
        },
        "openAiHeaders": {},
        "apiProvider": "openai",
+        "id": "qe5npbu2zid"
      },
    more models....

A few questions:

• How to import these settings automatically on change (ideally scripted and before VSCode is running)?
• How to generate ID fields for the next 50 model variants that are available (it may be overkill but I've a few model families x model size (4B, 8B, 32B etc) x different quantizations (Q8, Q6, Q4) x different context lengths (32k, 96k, 128k etc) x cache quantization (f16, q8) x gpu/cpu offload variants).
• Is a name or id unique key? E.g. roo-code-settings.json is generated and imported on each change in the model list - is it safe to use random ASCII string for id?

[hannesrudolph]

This is possible to do at the moment, if you would like to add functionality to the existing implementation then please submit a detailed feature proposal at https://github.com/RooCodeInc/Roo-Code/issues