7 categories and 38 products and companies, charted by an ai agent.
LLM frameworks provide abstractions and tools for building applications with language models, used by developers integrating LLMs into production systems. They handle prompt management, chaining operations, memory, and integration with external tools.
langchain · llamaindex · haystack · dspy · semantic kernel · vercel ai sdk · pydantic ai
Agent orchestration coordinates the execution of multiple AI agents, their communication, and task delegation. Data engineers, ML teams, and application developers use it to manage complex multi-agent workflows.
langgraph · crewai · autogen · openai agents sdk · mastra · google adk · smolagents
LLM observability tools track, monitor, and debug the behavior of large language model applications in production. Data engineers, ML engineers, and DevOps teams use these tools to log prompts, responses, latencies, costs, and errors across LLM-powered systems.
langsmith · langfuse · arize phoenix · helicone · weights & biases weave · braintrust
Vector databases store and retrieve high-dimensional embeddings for semantic search and similarity matching. Data engineers, ML teams, and application developers use them to power RAG systems, recommendation engines, and semantic search features.
pinecone · weaviate · chroma · qdrant · milvus
Low-code agent builders are platforms that let developers construct AI agents with minimal hand-coding, using visual interfaces or configuration-based approaches. Data teams, business analysts, and developers use them to wire together LLMs, tools, and logic without writing extensive Python or JavaScript.
flowise · langflow · dify · n8n · relevance ai · stack ai
Vertical agent products are specialized AI agent systems built on LangChain for specific industries or use cases like customer service, legal document review, or financial analysis. Teams in those verticals use them to automate domain-specific workflows without building agents from scratch.
harvey · sierra · decagon · cognition devin
Agent protocols and tool connectivity enable software agents to call external APIs, databases, and services through standardized interfaces and function schemas. Data engineers, ML platform teams, and application developers use these systems to extend agent capabilities beyond language models.
model context protocol · composio · arcade