| Setting | Default | Description |
|---|---|---|
eightbot.baseUrl |
https://api.8os.com |
Gateway URL |
eightbot.defaultModel |
auto |
Model used until you pick another |
eightbot.permissionMode |
default |
Permission mode at startup |
eightbot.reasoningEffort |
empty | Used when the model supports it |
eightbot.maxTurns |
200 |
Model turns per request |
eightbot.maxOutputTokens |
0 |
Output limit per turn. 0 uses up to 32,000 tokens or the model’s maximum |
eightbot.commandTimeoutSeconds |
120 |
Default command timeout |
eightbot.customInstructions |
empty | Added to every conversation |
eightbot.monthlySpendLimit |
0 |
Your own monthly limit in USD |
eightbot.pax8ClientId |
empty | Client account for live limits |
eightbot.planModel |
empty | Optional model used while in plan mode, for example a cheaper one |
eightbot.taskHint |
empty | Sent as X-Pax8-Task for routing rules |
eightbot.agentId |
8bot-vscode |
Sent as X-Pax8-Agent-ID |
eightbot.pax8McpServer |
false |
Offer the Pax8 MCP server to VS Code Chat |
All of these are user-level settings, so a workspace cannot override them.
For local development, the EIGHTOS_API_KEY environment variable is used when no key is stored.