xiry.

AI · Framework

AI Agents Stack (November 2024)

Market map categorizing the AI agent ecosystem across nine architectural layers

From the Xiry library · Updated 11 October 2026

This reference map catalogs roughly 60 key companies and open-source tools across the AI agent ecosystem as of late 2024. It organizes the landscape into distinct architectural layers, from foundational model serving and storage up to specialized vertical agents. It serves as a practical snapshot for identifying vendors and planning agent infrastructure.

The framework

AI Agents Stack — November 2024

VERTICAL AGENTS: Decagon, SIERRA, replit, perplexity, Harvey, MultiOn, Cognition, FACTORY, All Hands, Dosu, Lindy, 11x

AGENT HOSTING & SERVING: Letta, LangGraph, Assistants API, Agents API, Amazon Bedrock Agents, LiveKit Agents

OBSERVABILITY: LangSmith, arize, weave, Langfuse, AgentOps.ai, braintrust

AGENT FRAMEWORKS: Letta, LangGraph, AutoGen, LlamaIndex, CrewAI, DSPy, phidata, Semantic Kernel, AutoGPT

MEMORY: MemGPT, zep, LangMem, mem0

TOOL LIBRARIES: composio, Browserbase, exa

SANDBOXES: E2B, Modal

MODEL SERVING: vLLM, ollama, LM Studio, SGL, together.ai, Fireworks AI, groq, OpenAI, Anthropic, Mistral AI, Gemini

STORAGE: Chroma, Qdrant, milvus, Pinecone, Weaviate, NEON, Supabase

What it's for

  • Identify suitable tools and vendors for specific layers of an AI agent project
  • Compare architectural components when designing an autonomous agent system
  • Track and benchmark industry changes across the AI tooling ecosystem

Key points

  • Vertical Agents layer covering end-user specialized applications (e.g., Harvey, Cognition, Sierra)
  • Agent Hosting & Serving layer for deployment and orchestration runtimes
  • Observability and evaluation tools for tracking agent execution and outputs
  • Agent Frameworks layer comparing core development libraries (e.g., CrewAI, AutoGen, LangGraph)
  • Memory layer highlighting persistent context solutions (e.g., Mem0, Zep, MemGPT)
  • Supporting infrastructure layers including Tool Libraries, Code Sandboxes, Model Serving, and Storage

Pricing

Free

Details

Type
Framework
Category
AI
Topic
Market Maps
Open source
No
Platforms
other
For
AI engineers, technical product managers, and founders building AI agents
Tags
ai agents, market map, landscape, agent stack, frameworks, infrastructure

Why it was saved

  • @adamsilverman, 13 December 2024: A snapshot reference of the AI agents landscape in late 2024 — useful for finding vendors per layer or tracking how the stack evolves.

Related

Updated