Promptic

Promptic is a GenAI optimization platform that traces LLM apps with OpenTelemetry, benchmarks models and prompts on your data, and automatically tunes prompts, agents, and tool use for quality and cost.

Promptic aims to replace intuition-driven choices of prompts, models, and agent designs with a measured optimization loop. Teams upload use case data, choose the KPIs that matter such as quality, cost, and latency, and Promptic benchmarks candidate configurations on the same task, pinpoints where each wins or fails, and automatically tunes prompts, models, and tools to find the best-value configuration. The platform covers several layers of the GenAI stack: model selection across providers, automatic prompt improvement, agent architecture tuning, information extraction, and MCP tool selection accuracy. Tracing built on OpenTelemetry captures every LLM call, tool call, and step, turning production traces into optimization data. A coding agent integration lets tools like Claude Code run benchmarks, read failing traces, implement changes, and re-validate them. Promptic launched this version on Product Hunt on September 25, 2026. Plans start free with bring-your-own model keys and scale to team, business, and enterprise tiers priced in euros.