VibeAround v0.5.8: bridge OpenAI, Anthropic-style APIs, and more local coding agents

VibeAround v0.5.8 is live.

This release is a bigger rewrite of the Launch/Profile system, plus a more capable local API Bridge for coding agents.

The main idea: use the local coding agent you like, but route it to more model providers without constantly editing config files.

What changed:

- Local API Bridge: VibeAround can translate between OpenAI Responses, OpenAI Chat Completions, and Anthropic Messages.

- More providers: DeepSeek, Qwen/DashScope, Kimi, MiniMax, Z.AI/GLM, Gemini, OpenRouter, Azure OpenAI, and custom endpoints are easier to use from compatible local coding agents.

- Reworked Launch/Profile UI: profiles, API routes, model selection, terminal/workspace selection, and per-agent launch state are now cleaner.

- Better Claude Code / Codex API Bridge compatibility, including fixes for Qwen streaming and MiniMax reasoning output.

- Fixed channel-only onboarding for ACP agents, so Codex pickup from IM works more reliably.

- macOS Apple Silicon DMG is signed and notarized; Windows and Linux builds are available too.

Important note: this release includes a breaking change to profile configuration. If an old profile does not launch after upgrading, reopen its profile/connection settings, reselect the provider endpoint/API route/model, and save it again, or recreate the profile.

Release:

https://github.com/jazzenchen/VibeAround/releases/tag/v0.5.8

Curious to hear how people are mixing local coding agents with different model providers.