Skip to main content
  1. Sign up — Create an account at platform.respan.ai
  2. Create an API key — Generate one on the API keys page
  3. Add credits or a provider key — Add credits on the Credits page or connect your own provider key on the Integrations page

​
Prompt management

Instead of hardcoding prompts in your codebase, manage them on Respan. Every prompt has full version history: edit the text, model, temperature, or any configuration, commit a new version, and deploy it to production without touching your code. See the prompt quickstart for a hands-on walkthrough.
  • Manage prompts: create prompt templates, commit new versions, and compare changes across versions.
  • Variables: use {{variable_name}} to add dynamic content to your prompts. Variables let you reuse the same template across different inputs, and they map directly to testset columns when running experiments.
  • Use prompt in code: reference your prompt by ID in code. When you deploy a new version on the platform, your application picks it up immediately with no code changes.

​
AI gateway

The Respan AI gateway is a proxy that sits between your application and LLM providers. Route all your LLM traffic through a single API to get observability, reliability, and control without changing your SDK. See the gateway quickstart for a hands-on walkthrough.
  • Gateway setup: connect to 250+ models (OpenAI, Anthropic, Google, Azure, and more) through one unified API. Just change the base URL.
  • Fallbacks & retries: if your primary model fails, the gateway automatically retries on fallback models. Your users never see the error.
  • Load balancing: distribute traffic across multiple models or deployments to avoid rate limits and optimize cost.
  • Advanced configuration: enable caching to reduce cost and latency on identical requests, set rate limits, configure function calling, and more.
The gateway adds ~50-150ms of latency. If your product has strict latency requirements, use tracing for observability instead of routing through the gateway.