Core framework — agent lifecycle, action routing, tool registry, capabilities, streaming.

Multi-agent orchestration — groups, topologies, negotiation, workflows, debate, council, voting, auction, trust.

Vector database integrations for semantic search and RAG capabilities

Model Context Protocol (MCP) client and server — stdio, SSE, streamable HTTP transports, unified tool registry.

Built-in tool catalog — web, file, database, shell, code execution, ML, text, media, email, messaging.

Retrieval Augmented Generation (RAG) capabilities including document loading, text splitting, and embedding

Observability, security, validation — OpenTelemetry, Prometheus, prompt injection detection, sandboxing, invariants.

Agent-to-Agent (A2A) Protocol support for TnsAI framework - enables interoperability with other agent frameworks

Cognitive capabilities — FSM, planning, reasoning (ReAct, GoT, Causal), context management, RAG, knowledge bases.

LLM provider implementations — OpenAI, Anthropic, Google Gemini, Mistral, Ollama, AWS Bedrock, Cohere, Groq, HuggingFace, Azure OpenAI, OpenRouter, MiniMax, ZhipuAI.

Agent Client Protocol (ACP) support for IDE integration