v1.7.17

southwellmedia/velocityv1.7.17Aug 24, 2026by Ed1s0nZ

Upgrading?

AI Summary

This release disables automatic Eino checkpoint recovery to prevent stale responses and introduces detailed per-turn token usage tracking.

Key Highlights

  • Disabled automatic Eino checkpoint resume to prevent stale chat responses
  • Added per-turn token usage tracking including input, output, cached, and reasoning tokens
  • Added model call counts per conversation turn

New Features

  • Per-turn token usage tracking

Full Release Notes

这次修复关闭了聊天链路中的 Eino checkpoint 自动恢复,避免旧断点抢先恢复导致模型反复回复历史对话。

This update disables automatic Eino checkpoint resume in chat flows, preventing stale checkpoints from taking over and causing repeated old responses.

同时新增/完善了按轮次展示 Token 用量的统计能力,可查看每次对话的输入、输出、缓存、推理和模型调用次数。

It also adds/improves per-turn token usage tracking, showing input, output, cached, reasoning tokens, and model call counts for each conversation turn.

Related Repositories