OpenAI

Auto-detected#

Delimiter automatically detects requests to api.openai.com and reads rate-limit headers from every response. No setup needed beyond delimiter.init().

Supported headers#

OpenAI returns rate limit information on every API response:

x-ratelimit-limit-requests: 10000
x-ratelimit-remaining-requests: 7342
x-ratelimit-limit-tokens: 2000000
x-ratelimit-remaining-tokens: 1456000
x-ratelimit-reset-requests: 43s
x-ratelimit-reset-tokens: 12s

What gets parsed#

| Header | Parsed Field | | --- | --- | | x-ratelimit-limit-requests | requests_limit | | x-ratelimit-remaining-requests | requests_remaining | | x-ratelimit-limit-tokens | tokens_limit | | x-ratelimit-remaining-tokens | tokens_remaining | | x-ratelimit-reset-requests | reset_requests_ms (parsed from "43s" format) | | x-ratelimit-reset-tokens | reset_tokens_ms (parsed from "12s" format) |

Usage#

import { delimiter } from '@delimiter/sdk'
import OpenAI from 'openai'

delimiter.init('dlm_key')

// Use OpenAI as normal — Delimiter monitors automatically
const openai = new OpenAI({ apiKey: process.env.OPENAI_KEY })

await openai.chat.completions.create({ model: 'gpt-4o', messages: [...] })
await openai.embeddings.create({ model: 'text-embedding-3-small', input: '...' })

Works the same with LangChain, Vercel AI SDK, or raw fetch('https://api.openai.com/...') calls.

Rate limit tiers#

OpenAI rate limits vary by model and usage tier. Delimiter shows your actual limits as reported by the API — no configuration needed.