Repository navigation
Releases: ai1place/ai1place.github.io
Release list
Advisor v0.2.2
This is a pre-release of the Advisor macOS menu bar app. This offers a signed binary that can be used to safely install it onto your M-series Mac. v0.2.2 tweaked the menu bar icons.
Screenshot Example
The app itself uses AI to summarize work being performed by AI. You can opt into using Qwen locally, but it will attempt to use Apple Intelligence by default. Note that Apple Intelligence has fairly aggressive throttling limitations.
Tested agents
This works with multiple AI Agents out of the box:
| Agent | Status | Notes |
|---|---|---|
| Claude Code | Tested | Wired through Claude Code's own hooks. Confirmed on macOS and Windows. |
| Codex CLI | Tested | Wired through Codex's own hooks. Verified with a real codex exec run. Windows hooks are not separately verified. |
| Copilot CLI | Tested | Wired through Copilot's own hooks. The config includes both a bash and a powershell entry. Permission prompts are not yet tested. |
Cursor CLI (agent) |
Tested | Wired through Cursor's own hooks. Verified with interactive agent sessions and agent -p. MCP events are not yet tested. |
| Grok Build | Tested | Wired through Grok Build's own hooks. Verified with a real grok -p run. Confirmed on macOS and Windows. |
Limitations
- While Claude Code works, Claude the app, and
claude.aido not fire hooks on your local machine so there is no way to show them. - Codex and ChatGPT has the same limitation. Codex fires fewer hooks.
- Excluding Grok, all of the supported AI agents will fire questions, but none of them currently report the questions have been answered so Advisor will report a question state until another hook fires for the same chat session.
- Advisor can properly bring forward VS Code and IntelliJ, but it cannot select the tab with the agent conversation.
Advisor v0.2.1
This is a pre-release of the Advisor macOS menu bar app. This offers a signed binary that can be used to safely install it onto your M-series Mac.
Screenshot Example
The app itself uses AI to summarize work being performed by AI. You can opt into using Qwen locally, but it will attempt to use Apple Intelligence by default. Note that Apple Intelligence has fairly aggressive throttling limitations.
Tested agents
This works with multiple AI Agents out of the box:
| Agent | Status | Notes |
|---|---|---|
| Claude Code | Tested | Wired through Claude Code's own hooks. Confirmed on macOS and Windows. |
| Codex CLI | Tested | Wired through Codex's own hooks. Verified with a real codex exec run. Windows hooks are not separately verified. |
| Copilot CLI | Tested | Wired through Copilot's own hooks. The config includes both a bash and a powershell entry. Permission prompts are not yet tested. |
Cursor CLI (agent) |
Tested | Wired through Cursor's own hooks. Verified with interactive agent sessions and agent -p. MCP events are not yet tested. |
| Grok Build | Tested | Wired through Grok Build's own hooks. Verified with a real grok -p run. Confirmed on macOS and Windows. |
Limitations
- While Claude Code works, Claude the app, and
claude.aido not fire hooks on your local machine so there is no way to show them. - Codex and ChatGPT has the same limitation. Codex fires fewer hooks.
- Excluding Grok, all of the supported AI agents will fire questions, but none of them currently report the questions have been answered so Advisor will report a question state until another hook fires for the same chat session.
- Advisor can properly bring forward VS Code and IntelliJ, but it cannot select the tab with the agent conversation.