mirror of https://github.com/jlengrand/Maestro.git synced 2026-03-10 08:31:19 +00:00

Files

Pedram Amini c30c053e0f feat: add usage stats tracking, UI polish, and LLM settings feature flag

- Add real-time token usage and cost tracking from Claude Code responses
  - New UsageStats type with tokens, cost, and context window
  - Context window tooltip shows detailed token breakdown
  - Cost displayed in main panel header and session list

- Improve input area UX with $ prefix for terminal mode

- Add Cmd+L shortcut to insert markdown checkbox in Scratchpad

- Add terminalOnly flag for slash commands (/jump is terminal-only)

- Disable LLM Settings panel behind feature flag (documented in BACKBURNER.md)

- Fix AgentSessionsBrowser to use layer stack for Escape handling

- Update README with cost tracking feature and keyboard shortcut
- Update CLAUDE.md with process:usage event and usageStats session field

2025-11-26 01:42:30 -06:00

1.6 KiB

Raw Blame History

Backburner Features

This document tracks dormant features that are implemented but disabled via feature flags. These features may be re-enabled in future releases.

LLM Settings Panel

Status: Disabled Feature Flag: FEATURE_FLAGS.LLM_SETTINGS in src/renderer/components/SettingsModal.tsx Disabled Date: 2024-11-26

Description

The LLM Settings panel provides configuration options for connecting to various LLM providers directly from Maestro. This feature was designed to enable a built-in AI assistant for the scratchpad or other future AI-powered features within the application.

Supported Providers

OpenRouter - API proxy supporting multiple models
Anthropic - Direct Claude API access
Ollama - Local LLM inference

Configuration Options

LLM Provider selection
Model slug/identifier
API key (stored locally)
Connection test functionality

Files Involved

src/renderer/components/SettingsModal.tsx - Main settings UI with LLM tab
Settings stored in electron-store: llmProvider, modelSlug, apiKey

Re-enabling

To re-enable this feature:

Open src/renderer/components/SettingsModal.tsx
Find the FEATURE_FLAGS constant at the top of the file
Set LLM_SETTINGS: true

const FEATURE_FLAGS = {
  LLM_SETTINGS: true,  // LLM provider configuration (OpenRouter, Anthropic, Ollama)
};

Reason for Disabling

Currently not in use as Maestro focuses on managing external AI coding agents (Claude Code, etc.) rather than providing built-in LLM functionality. May be re-enabled when there's a use case for direct LLM integration within Maestro.

1.6 KiB Raw Blame History