Releases: johnbean393/Hivecrew
Releases · johnbean393/Hivecrew
1.25
Update
New Features
- Switch to Hivecrew Office skills
- Parallel image generation
- Support
gpt-image-2 - Update VM template to v0.0.20
- Hivecrew agent pointer
- Enhance Hivelink iOS integration
- Hivelink continue from task
- Hivelink ChatGPT OAuth voice provider
- Hivelink voice picker
- OpenAI OAuth realtime voice
- Support OpenAI OAuth image generation
Bug Fixes
- Raise image generation timeout
- Worker model reset
- Nudge agent to place deliverables in outbox
- Remove broken AX traversal tool
- Delete VM template before update
- Sync OpenAI VAD parameters
- Hivelink improve OpenAI VAD
- Hivelink persist model choice
- Automatically use OAuth for voice
- Remote origin task state
- Hivelink background refresh crash
Requirements
- macOS Sequoia (15.0) or later
- Apple Silicon Mac (M1 or newer)
- 16GB RAM recommended for concurrent agents
- ~64GB free disk space per VM template
1.24
Update
New Features
- Detached task trace windows
- Increase agent step limit
Bug Fixes
- complete remaining translations
- Add missing translations
- Remove minimum task time
- Revoke tunnels properly
- Hivelink cluster state
- Hivelink honor prompt bar settings
- Update worker URL
- Hivelink delete account
Requirements
- macOS Sequoia (15.0) or later
- Apple Silicon Mac (M1 or newer)
- 16GB RAM recommended for concurrent agents
- ~64GB free disk space per VM template
1.23
Update
New Features
- Hivelink bulk deliverable actions
- Privacy and terms
- PrivacyInfo declarations
- Hivelink delete account
- Hivelink proximity sensor
- Hivelink inbound calls
- Hivelink notifications
- Hivelink plan rendering
- Hivelink system integrations
- Hivelink CallKit
- Hivelink tasks
- Task list and prompt bar UI
- Hivelink cluster UI
- Hivelink authentication
- Hivelink project setup
- Add HivecrewCore
- Add Hivelink
- Screen sharing HUD
Bug Fixes
- Hivelink deafening volume
- VoIP pushes
- Hivelink end call
- Dynamic health checks
- Remove tunnel self pings
- Hivelink remove "All Tasks Done" event
- Hivelink landscape layout
- Missing VoIP pushes
- Hivelink duplicate calls
- Hivelink notification spam
- Hivelink inbound calls
- Hivelink screen broadcast
- Hivelink task trace
- Hivelink plan mode
- Hivelink plan mode
- Dynamic notch transcript
- Screen capturing animation
- DuckDuckGo search
- Remote task continuation
- Show update completion
Requirements
- macOS Sequoia (15.0) or later
- Apple Silicon Mac (M1 or newer)
- 16GB RAM recommended for concurrent agents
- ~64GB free disk space per VM template
1.22
Update
New Features
No new features in this release.
Bug Fixes
- Template OOS syntax error
- Template update OOS error
Requirements
- macOS Sequoia (15.0) or later
- Apple Silicon Mac (M1 or newer)
- 16GB RAM recommended for concurrent agents
- ~64GB free disk space per VM template
1.21
Update
New Features
- Nudge to produce PDFs
Bug Fixes
- Add missing translations
- Voice mode robustness
- LLM provider persistence
- Remote execution state
- Voice transcript auto-scrolling
Requirements
- macOS Sequoia (15.0) or later
- Apple Silicon Mac (M1 or newer)
- 16GB RAM recommended for concurrent agents
- ~64GB free disk space per VM template
1.20
Update
New Features
- Granular token counter
- Decentralized clustering
- Voice isolation onboarding
- Improve cluster UI
- Improve cluster scheduling
Bug Fixes
- Remote step count
- Improve voice mode anti-interruption
- Add missing translations
Requirements
- macOS Sequoia (15.0) or later
- Apple Silicon Mac (M1 or newer)
- 16GB RAM recommended for concurrent agents
- ~64GB free disk space per VM template
1.19
Update
New Features
- Edit and rerun task
- Cluster task dispatch
- Support OpenAI voice provider
- Voice agent read file tool
- Visualize voice mode tool calls
- Voice agent attachment search
- Voice mode
Bug Fixes
- Add missing translations
- De-clutter settings
- OpenAI Codex new response shape
- Get rid of worker roles
- Voice mode token counter
- Continuity camera
- Voice agent interruption
- Remove artifacts on failed VM download
- Improve subagent token efficiency
Requirements
- macOS Sequoia (15.0) or later
- Apple Silicon Mac (M1 or newer)
- 16GB RAM recommended for concurrent agents
- ~64GB free disk space per VM template
1.18
Update
New Features
No new features in this release.
Bug Fixes
- Subagent errors
- Subagent timeout error
- Extract artifacts on fail
- Prevent VM sleep
- Subagent failure
Requirements
- macOS Sequoia (15.0) or later
- Apple Silicon Mac (M1 or newer)
- 16GB RAM recommended for concurrent agents
- ~64GB free disk space per VM template
1.17
Update
New Features
- Persist Codex plan name
- Codex limits
- Show token usage
- Persist provider level selected model
- Agent read / write to host FS
- Bumped VM template
Bug Fixes
web_searchfallback chain- Orphaned VM deletion
- Token usage w/ subagents
- OpenRouter vision parsing
- Encourage edit in place
Requirements
- macOS Sequoia (15.0) or later
- Apple Silicon Mac (M1 or newer)
- 16GB RAM recommended for concurrent agents
- ~64GB free disk space per VM template
1.16
Update
New Features
- Continue from task
- Default to Responses API
- Improve skills
- Support Codex Fast mode
- Add ChatGPT sign in to onboarding
- Rename task
- Web UI multi model tasks
- Reasoning effort in web UI
- Dynamic LLM call timeouts
- Sort models by version number
- Reasoning effort controls
- Support /v1/responses and ChatGPT OAuth providers
- Rerun with model
- Improve multi model task
- Add VM template update menu check
- Update VM template
Bug Fixes
- Context compaction with responses API
- Webpage extraction chain
- Codex subagent model slug
- ChatGPT sign in caveat
- Codex streaming only
- Web UI default reasoning effort
- Session trace control crowding
- Add sheet exit buttons
- Rerun sheet height
- Default reasoning on web UI
- Multi-model task init
- Reasoning effort picker for multi-model
- Added missing translations
- Fixed notification permission prompt timing
- Fixed credential logging verbosity
- Fixed imported template version metadata
- Fixed ask later prompt delay
- VM template check
Requirements
- macOS Sequoia (15.0) or later
- Apple Silicon Mac (M1 or newer)
- 16GB RAM recommended for concurrent agents
- ~64GB free disk space per VM template