CheckOps home

Tools directory

R

CheckOps

Respan

Respan is an open-source AI gateway with built-in observability and evals that routes calls to 1,000+ models through one API key and traces every call, with paid tiers for logs and scores.

Visit website

Latest public signals

Not refreshed yet. Refresh reads the public site and, when the entry has a GitHub repository, the latest release.

Visit Website

Introduction

Respan is an AI gateway first: one API key for over 1,000 models, OpenAI-compatible endpoints and claimed token discounts. Every call through it is traced, and its YC profile describes it as an open-source, extensible gateway with observability and evals.

Setup is a base-URL change in the OpenAI, Vercel AI SDK or LangChain clients, or the @respan/cli npm package. Evaluation scores are billed per thousand, so evals sit alongside logging rather than in a separate tool.

Launch and background: Y Combinator’s company directory lists Respan in its Winter 2024 batch, with the one-liner “Open-source, extensible AI gateway with observability and evals”.

Features

  1. Gateway: 1,000+ models behind one key.
  2. Tracing: Every call logged.
  3. Scores: Evaluation results on calls.
  4. Drop-in: Change the base URL.
  5. CLI: respan auth login.

Pros

  1. Free to start: No card needed.

Cons

  1. Gateway dependency: Traffic must route through Respan.
  2. Eval depth: Lighter than dedicated eval platforms.
npm install -g @respan/cli
respan auth login

Then point your OpenAI-compatible client at https://gateway.respan.ai/v1/.

Resources:

Pricing

Respan starts free with no card, and lists a $199 a month plan, usage rates such as $8 per 100k logs and $1 per 1k scores, $15 per member and a $500 a month add-on.

  1. Teams wanting routing and observability in one hop.

Conclusion

Respan suits teams that want a gateway and basic evals without extra plumbing.

Use a dedicated eval tool for deep testing.

On CheckOps, Respan is filed under the Production monitoring & observability stage. Related tools listed here include Helicone, Langfuse, LangSmith, OpenLLMetry, Braintrust and Arize Phoenix. A short trial on a real project, with your own code, data and pipeline, is usually the fastest way to choose between them, because fit matters more than feature lists. Facts here come from the official site, documentation and repository, checked on 11 November 2026.

Alternatives

Other tools on CheckOps that cover similar ground. Open a page, or compare one with Respan side by side.