What's Changed
🌄 Highlights
🤖 Agents
- Introducing LibreChat's No-Code Agentic Framework (learn more)
- Create custom AI assistants with supported providers, similar to Assistants API and ChatGPT's GPTs
- Supports Anthropic, OpenAI, AWS Bedrock, Azure OpenAI, and custom endpoints
- Enhanced with tools like Code Interpreter, DALL-E-3, Tavily Search, and more
- Dynamic tool creation through OpenAPI specs and MCP Servers (#4627)
🎯 Code Interpreter API
- Execute code securely in multiple languages through a powerful sandbox environment (learn more)
- Seamless file handling and management with secure session-based organization (#4860)
🔌 MCP Support (Tools)
- Introducing Model Context Protocol integration
- the "USB-C of AI" that standardizes how applications provide context to LLMs (#5015)
- Learn more about MCP and LibreChat integration at modelcontextprotocol.io
🚀 New Models Support
- Gemini 2.0 and Flash Thinking Mode (#4951, more info)
- New o1 "Full" Model and o1-preview/mini streaming support (#5039), (#4760)
- Claude 3.5 Haiku (#4629)
- AWS Nova Models (#4852)
See detailed release notes at https://www.librechat.ai/changelog/v0.7.6
New Contributors
- @timmanik made their first contribution in #4494
- @MSITETOP made their first contribution in #4486
- @senk made their first contribution in #4763
- @siddharthsambharia-portkey made their first contribution in #4625
- @wipash made their first contribution in #4787
- @leon-juenemann made their first contribution in #4659
- @dennis531 made their first contribution in #4506
- @thingersoft made their first contribution in #4822
- @philwilliammee made their first contribution in #4896
- @thelinuxkid made their first contribution in #4958
- @rio2dev made their first contribution in #4881
- @alex-torregrosa made their first contribution in #4920
Full Changelog: v0.7.5...v0.7.6