Crail now tracks 146 vendors across 23 categories
We started with two deep categories (AI coding agents, LLM & agent infrastructure) and have expanded to cover the broader agent/LLM/generative-AI tool landscape, plus classic high-volume SaaS categories.
Crail launched tracking two categories in depth — AI coding agents and LLM & agent infrastructure — on the theory that these were the clearest early wedge for an agent-native software directory. We’ve since expanded to 23 categories and 146 vendors, covering both the wider agent/generative-AI tool landscape and a set of classic high-volume B2B SaaS categories.
What’s new
Beyond the original 10 categories, Crail now tracks:
- AI Agent Frameworks & Orchestration (LangChain, LangGraph, CrewAI, AG2, LlamaIndex, Semantic Kernel)
- Vector Databases & RAG Infrastructure (Pinecone, Weaviate, Qdrant, Chroma, Zilliz, MongoDB Atlas Vector Search)
- AI Model Hosting & Inference (Together AI, Fireworks, Replicate, Groq, Modal, Baseten)
- AI Voice & Speech, Generative Image AI, Generative Video AI
- AI Search & Retrieval APIs, AI Customer-Facing Support Agents, AI Sales & SDR Agents
- Synthetic Data & Data Labeling, AI Meeting Assistants, AI Browser & Computer-Use Agents, AI Writing & Content Assistants
What stays the same
Every vendor record — old category or new — gets the same treatment: pricing verified against the vendor’s own pages, an agent-readiness score across the same six sub-dimensions, MCP support checked (not just claimed), and a lastVerifiedAt date on every field group. No vendor payment influences ranking in any category, old or new — see our methodology.
Browse the full list on the categories page.