Langfuse launches AI Gateway

Langfuse has launched its AI Gateway in public beta as a centralized, open-source entry point for organizations to route, govern, and monitor AI model traffic across applications, coding agents, assistants, and internal tools. The gateway records usage, costs, identities, and optionally full prompts and responses as Langfuse traces, allowing teams to analyze spending by model, provider, team, or cost center while using existing tracing, evaluation, annotation, and search capabilities. Designed to address growing enterprise needs around AI cost control, security, compliance, and visibility, it supports centralized model approvals, configurable data logging, spend limits, exports to data lakes, and integration with OpenTelemetry and other Langfuse integrations. Built in Rust, the gateway is MIT-licensed, can be self-hosted as a lightweight Docker image or used through Langfuse Cloud, and is intended to preserve customers’ direct provider contracts and compliance terms. The beta initially supports OpenAI and Anthropic, requires only a base URL and API key change, and plans to add providers, routing, fallbacks, guardrails, PII redaction, and enterprise access-management features.

Use this post, company, and trend context to find content marketing opportunities, perform competitive analysis, or address product feature gaps via the Plushcap MCP server or the Plushcap API.