Foglamp

Foglamp shows what AI agents built on the Vercel AI SDK are doing: cost, latency, tokens, traces, evals and alerts for every generateText and streamText call, added with two lines of code. Its dashboard follows every model and tool call with prompts and responses in one place, tracks spending by model, agent and customer, and can be self-hosted or used as a hosted service. It suits developers shipping AI features who need to know what agents cost and when answers degrade.