Changelog

​

Recent Releases

​

v0.1.4.post3 (2026-02-28)

• Cleaner context: Improved prompt building
• Hardened session history: More reliable conversation tracking
• Smarter agent: Better decision-making
• Security improvements
• Bug fixes and stability improvements

​

v0.1.4.post2 (2026-02-24)

• Redesigned heartbeat system: More reliable periodic task execution
• Prompt cache optimization: Faster responses with caching
• Hardened provider stability: Better error handling
• Hardened channel stability: Improved connection resilience

• Virtual tool-call heartbeat
• Slack mrkdwn formatting fixes
• Discord typing indicator fixes

​

v0.1.4.post1 (2026-02-21)

• VolcEngine support
• MCP custom auth headers
• Anthropic prompt caching

• Feishu multimodal file handling
• Slack file sending
• Discord long message splitting

• More reliable memory subsystem
• Subagents work in CLI mode
• Session poisoning fix
• WhatsApp deduplication
• Windows path compatibility
• Mistral provider compatibility

​

v0.1.4 (2026-02-17)

• MCP integration: Model Context Protocol support
• Progress streaming: Real-time updates during agent execution
• New providers: Multiple new LLM provider integrations
• Channel improvements: Better stability across all channels

• ClawHub skill search and install
• OpenAI Codex provider with OAuth
• MiniMax support

• Enhanced CLI experience
• Better session management
• Auto workspace template sync

​

v0.1.3.post7 (2026-02-13)

• Security hardening across all components
• Important: Upgrade recommended to address security issues

• Redesigned memory system: Less code, more reliable
• Better error handling
• Stability improvements

​

v0.1.3.post6 (2026-02-10)

• Slack support
• Email channel support
• QQ channel support

• Enhanced CLI experience
• Better provider detection
• Improved error messages

​

v0.1.3.post5 (2026-02-07)

• Qwen/DashScope provider support
• Refactored provider system: 2-step provider addition

• Better provider registry
• Simplified provider configuration
• Improved documentation

​

v0.1.3.post4 (2026-02-04)

• Multi-provider support: Switch between LLMs easily
• Docker support: Containerized deployment
• DeepSeek provider
• Moonshot/Kimi provider
• Discord channel integration

• Enhanced security hardening
• Better configuration validation

​

Earlier Releases

​

v0.1.3 Series

• Added 10 chat channels (Telegram, Discord, WhatsApp, etc.)
• Implemented provider registry pattern
• Enhanced security features
• Docker support
• MCP integration

​

v0.1.2 Series

• Stable agent loop
• Multiple built-in tools
• Session management
• Memory system

​

v0.1.1 Series

• Telegram channel
• Basic tool set (shell, filesystem, web)
• Configuration system

​

v0.1.0 (2026-02-02)

• Core agent loop
• Telegram channel
• Basic tools (exec, read, write, web_search)
• OpenRouter provider
• CLI interface
• Configuration via config.json

​

Version History Summary

​

Breaking Changes

​

v0.1.4.post3

• Before: Empty allowFrom = allow all senders
• After: Empty allowFrom = deny all senders (use ["*"] for allow-all)
• Migration: Add "allowFrom": ["*"] to channel configs if you want to allow all users

​

v0.1.3.post7

• Old memory format no longer supported
• Migration: Manually copy important facts to new MEMORY.md format

​

Upgrade Guide

​

From v0.1.4.postX to latest

pip install --upgrade nanobot-ai

​

From v0.1.3.x to v0.1.4.x

• MCP server config format (compatible, no action needed)
• Progress streaming metadata (internal, no action needed)

1. Upgrade: pip install --upgrade nanobot-ai
2. Test: nanobot agent -m "Hello"

​

From v0.1.2.x to v0.1.3.x

• Provider registry pattern (config format unchanged)
• Memory system redesigned

1. Backup config: cp ~/.nanobot/config.json ~/.nanobot/config.json.backup
2. Upgrade: pip install --upgrade nanobot-ai
3. Migrate memory: Review ~/.nanobot/workspace/MEMORY.md
4. Test: nanobot agent -m "Hello"

​

Deprecation Notices

​

Current

​

Future

• Old config format (pre-v0.1.3) will no longer be supported
• Migration tool will be provided

​

Download Statistics

​

Contributing to Releases

​

Release Process

1. Version bump: Update version in pyproject.toml
2. Changelog: Update CHANGELOG.md (if exists)
3. Tag: Create git tag git tag v0.1.x
4. Build: python -m build
5. Publish: twine upload dist/*
6. GitHub Release: Create release with notes

​

Semantic Versioning

• Major (v1.0.0): Breaking changes, major rewrites
• Minor (v0.2.0): New features, non-breaking changes
• Patch (v0.1.5): Bug fixes, minor improvements
• Post (v0.1.4.post1): Hotfixes, urgent patches

​

Release Notes Guidelines

1. Title: Short, descriptive (e.g., “MCP Support & Progress Streaming”)
2. Summary: 2-3 sentences explaining the release
3. Sections:
   • New Features: What’s new
   • Improvements: What’s better
   • Bug Fixes: What’s fixed
   • Breaking Changes: What breaks (if any)
4. Credits: Thank contributors
5. Links: Link to PRs, issues, discussions

​

Stay Updated

• GitHub Releases: Watch the repo
• Discord: Join announcements channel
• PyPI: Subscribe to updates

​

Frequently Asked Questions

​

How often are releases?

• Patch releases: As needed (bug fixes)
• Minor releases: Every 1-2 weeks (new features)
• Major releases: When breaking changes occur

​

Should I always upgrade?

• Security releases: Yes, immediately
• Feature releases: When you need the features
• Patch releases: Recommended (usually safe)

​

How do I rollback?

pip install nanobot-ai==0.1.4.post2  # Specific version

​

What if an upgrade breaks my setup?

1. Rollback to previous version
2. Check changelog for breaking changes
3. Open a GitHub issue with details
4. We’ll help you migrate or fix the issue

​

Version Support

​

Currently Supported

• v0.1.4.x: Active development, bug fixes
• v0.1.3.x: Security fixes only

​

End of Life

• v0.1.2.x: No longer supported
• v0.1.1.x: No longer supported
• v0.1.0: No longer supported

​

Long-Term Support (Future)

• LTS versions: 12 months of support
• Regular versions: 3 months of support