platoseedplatoseed
Self-driving observability, evals, and gateway for AI agents
Respan (formerly Keywords AI) gives teams a unified control plane to trace and evaluate agent behavior without guesswork, automatically surface issues, and fix what breaks faster. Respan connects production observability, automated and human-in-the-loop evaluations, and an adaptive AI gateway to close the loop between detection, decision, and action - so agents improve continuously in production. Respan is trusted by 100+ AI startups and enterprise teams. Today, the platform process 1B+ logs and 2T+ tokens every month, supporting more than 6.5M end users.
Respan is an LLM engineering platform that unifies observability, evals, prompt optimization, and a unified gateway for AI agents. It enables tracing, evaluating, and deploying AI-powered workflows from a single interface with model fallbacks and cost controls.
Respan provides a gateway that routes API calls to 500+ models, with a single API or provider-native SDK passthrough. It logs every request, offers model fallbacks and retry/backoff, spend limits per API key with alerts, and dashboards to monitor usage, cost, latency, and errors. It supports building and executing evaluation workflows (human review + AI judges) tied to real production traces, and allows prompt/model/workflow promotion from the UI with version control and rollout logic. It also enables tracing and observability of every prompt, tool call, and response, plus online evals on live production data and dataset creation from production traces for experiments across variants.
Who it’s for: Engineering teams shipping AI-powered products that rely on multiple LLM providers and need observability, evaluation, and governance for production AI agents.
Funding/capitalization not explicitly stated; active product with customer testimonials and multiple integrations; YC mention in site navigation implies startup traction.
CEO @ Respan (formerly Keywords AI) - observability, evaluations, and AI gateway for agents World-class AI teams use Respan to: - trace and evaluate agent behavior without guesswork - automatically surface issues - fix what breaks, faster Previously @ UIUC (dropped out) and Apple Product Design
CTO @ Respan (formerly Keywords AI) - the developer platform to for agent observability & evals.
Dead simple way to deploy & monitor AI apps with 2 lines of code.
Keywords AI provides infrastructure, tooling, and a collaborative workflow to build, deploy, monitor, and fine-tune LLM applications. It offers OpenAI-compatible deployment with 100+ models, a prompt manager, request logging, performance evaluation, usage analytics, data curation, and one-click fine-tuning for developers.
From the original launch (Apr 2024) — may be outdated.
Formerly “Keywords AI” · why startups rename →
AI DevOps Agents
Understand why your AI agent breaks. Iterate fast to fix it.
