# AI infrastructure: the tools companies use

1,069 companies on usedby use a tool of its 6 categories with a public customer. Each category ranks its own tools: tools of different jobs are not ranked against each other. Counts are of companies on usedby, not of the whole market. Latest check 2026-10-14: 1,385 of 1,436 public lines checked.

> What engineering teams use to build, run and monitor AI products: inference clouds, frameworks, observability, ML platforms, vector databases and speech APIs.

## Categories

- [Vector databases & search engines](https://www.usedby.ai/tools/category/vector-databases.md): 290 companies, 9 tools with a public customer; most used: Qdrant
- [ML platforms & training data](https://www.usedby.ai/tools/category/ml-ops-training.md): 256 companies, 12 tools with a public customer; most used: Weights & Biases
- [AI inference & GPU cloud](https://www.usedby.ai/tools/category/llm-inference.md): 230 companies, 10 tools with a public customer; most used: Baseten
- [Agent frameworks & AI gateways](https://www.usedby.ai/tools/category/llm-ops-frameworks.md): 188 companies, 13 tools with a public customer; most used: LangChain
- [Speech & voice APIs](https://www.usedby.ai/tools/category/speech-audio-api.md): 186 companies, 5 tools with a public customer; most used: Recall.ai
- [LLM observability & evals](https://www.usedby.ai/tools/category/llm-observability-evals.md): 153 companies, 6 tools with a public customer; most used: LangSmith

---
Source: https://www.usedby.ai/tools/category/ai-infrastructure · How we check: https://www.usedby.ai/methodology