LibreFang 0.6.0 Released

We're thrilled to release LibreFang v0.6.0—packed with enterprise-grade features, expanded LLM provider support, and a more powerful API for building intelligent agents. Here's what changed.

🧠 Memory & Context Intelligence

LibreFang now ships with mem0-style proactive memory, letting agents learn and evolve from past conversations. We've also significantly improved the context engine—it's now more accurate, more resilient, and ships with a plugin management system that lets you customize how agents retrieve and reason over context. For advanced use cases, you can register multiple custom plugin registries.

🤖 LLM Provider Ecosystem

This release dramatically expands LLM flexibility:

• NVIDIA NIM is now a first-class provider—ideal for organizations running local or on-prem inference
• Vertex AI joins the lineup with full OAuth2 authentication support
• Intelligent LLM fallback: Web search key rotation, data-driven routing, and health-aware LLM selection ensure your agents gracefully degrade if a provider goes down

📡 Channels & Real-Time Messaging

Agents can now reach more users through richer channels:

• Telegram streaming with progressive message updates (no more "typing..." delays)
• Multimedia support for Telegram and Discord—images, video, audio all work seamlessly
• Improved dispatch reliability—silent message drops are a thing of the past
• Sender identity propagation—channels now forward the original sender's identity to agent context, enabling personalized responses

⚙️ Workflows & Automation Overhaul

Workflows are now deeply integrated with your infrastructure:

• Cron-triggered workflows: Schedule agents to wake up on a timer and execute tasks automatically
• Auto-registration at startup: Local workflow definitions load without manual wiring
• New workflow REST endpoints: Full CRUD operations on workflows with /api/workflows/:id

📊 API Expansion

We added 8 new REST endpoints to give you finer-grained control:

• GET /api/providers/:name, GET /api/workflows/:id, GET /api/channels/:name, GET /api/cron/jobs/:id, GET /api/mcp_servers/:name
• PUT/DELETE /api/workflows/:id for workflow lifecycle management
• DELETE /api/agents/:id/files/:filename for file cleanup
• Agent list improvements: filtering, pagination, and sorting for handling thousands of agents

🌐 Infrastructure & Developer Experience

• HTTP proxy support for all outbound connections (essential for enterprise environments)
• Docker image now bundles Python and Node.js runtimes—no more installing dependencies after pulling the image
• Streaming improvements: Better handling of chunked responses from thinking models (Gemini's o1 style)
• SSL/TLS reliability: Falls back to bundled Mozilla CA roots if system certificates are unavailable

🔧 Bug Fixes & Maintenance

• Fixed Slack thread detection for proper conversation threading
• Corrected MIME type handling for vision models on images
• SHA-256 Nostr pubkey derivation for security compliance
• 10 upstream parity bug fixes to stay in sync with the broader ecosystem
• Full validation of agent list operations

Install / Upgrade

# Binary
curl -fsSL https://get.librefang.ai | sh

# Rust SDK
cargo add librefang

# JavaScript SDK
npm install @librefang/sdk

# Python SDK
pip install librefang-sdk

Links

• Full Changelog
• GitHub Release
• GitHub
• Discord
• Contributing Guide