PatternsAgent Patterns

OpenAI Prompt Caching (AI Gateway)

Lightweight ToolLoopAgent pattern for OpenAI prompt caching through Vercel AI Gateway: stable system instructions, promptCacheKey, and per-turn cachedPromptTokens metrics.

Loading preview…

Problems solved

Technical challenges this implementation handles out of the box.

Enable OpenAI prompt caching with promptCacheKey via AI Gateway
Reuse a stable system prefix across ToolLoopAgent turns
Inspect cachedPromptTokens from provider metadata

Use cases

Products and workflows this pattern is designed to support.

Lowering cost on repeated agent system prompts
Teaching promptCacheKey grouping with ToolLoopAgent
Comparing cache metrics across multi-turn sessions

Setup

Requirements, wiring steps, and what this pattern adds to your project.

Getting started

Pick how you want to pull this pattern in — then wire env vars and routes the same way.

1
Install from the preview toolbar
Copy the install command above and run it in your project — Pro patterns include a short-lived token.
2
Add environment variables to .env.local
AI_GATEWAY_API_KEY
3
Point client fetches at your API routes
Replace preview `/view/.../api/...` paths with `/api/...` in the installed files.
4
Customize the agent and tool files
Adapt prompts, tools, and stop conditions for your product — Copy for AI in the toolbar helps seed that work.
5
Run your dev server and open the pattern route
Install dependencies if needed, then start the app and verify responses.

Environment variables

AI_GATEWAY_API_KEYGet key

Capabilities

AI SDK APIs

ToolLoopAgentgatewayproviderOptionsstepCountIs

Providers

OpenAI

Files added

8 files

app/page.tsx
app/layout.tsx
app/api/cache-openai/route.ts
lib/cache-openai-agent.ts
lib/cache-metrics.ts
lib/stable-system-prompt.ts
lib/request-utils.ts
README.md

Dependencies

9 total

npm packages

AISDKlucide-reactNext.jsReactZod

Registry components

alertbadgebuttonskeleton

Critical files

lib/cache-openai-agent.ts
ToolLoopAgent with gateway model, stable instructions, and providerOptions.openai.promptCacheKey.
cache-openai/route.ts
Runs agent.generate({ messages }) and returns text plus cachedPromptTokens metrics.