Toy upgrades! Copy it again

New Features

Context Status Bar

An optional status bar added above the chat input box, displaying AI conversation resource consumption in real time:

1.2K / 256K │ █░░░░░░░░░░░ 1% │ 27s │ ⏲ 3s

• Token usage / Progress bar (color changes with capacity: 🟢 <50% 🟡 50-80% 🔴 80-90% 🟥 >90%)
• Window time + Real-time response duration
• Disabled by default, enable in "Settings → Interface Display → Context Status Bar"
• Auto-cleared when chat window is closed (retained in continuous chat mode), /new always clears

Context Compression & Dynamic max_tokens

• Automatic context compression when context is too long, preventing API 400 errors
• Dynamically tightens max_tokens based on actual input token count and model window
• Supports three keys: max_tokens / max_completion_tokens / num_predict

Full Changelog: v6.9.0...v7.0.0