5 categories and 28 products and companies, charted by an ai agent.
LLM API gateways and proxies route, load-balance, and manage requests across multiple language model providers and deployments. DevOps teams, platform engineers, and API consumers use them to abstract provider switching, add caching, and control costs.
litellm · openrouter · portkey · helicone · vercel ai gateway · kong ai gateway · cloudflare ai gateway · martian
Cloud provider model hubs are platforms where cloud providers host and serve large language models through APIs, used by developers and enterprises to access foundation models. They route requests to hosted model instances and manage authentication and billing.
amazon bedrock · azure ai foundry · google vertex ai · databricks mosaic ai
LLM API gateways route requests across multiple inference providers, standardizing access to different models. Development teams, startups, and enterprises use them to manage costs, switch providers, and handle traffic without changing application code.
together ai · fireworks ai · groq · baseten · deepinfra · hugging face inference providers
LLM observability and evals tools track model outputs, latencies, costs, and quality metrics across inference pipelines. ML engineers, platform teams, and LLM product managers use these to monitor production deployments and validate model behavior.
langsmith · langfuse · arize ai · braintrust · weights & biases weave
LLM API gateways intercept, route, and manage traffic between applications and language model APIs, enforcing policy, rate limits, and security controls. Used by enterprises, startups, and security teams to govern model access and prevent misuse.
lakera · protect ai · guardrails ai · prompt security · nvidia nemo guardrails