LumeAPIText & chat
GPT, Claude, Gemini, DeepSeek, and more through OpenAI-style Chat Completions — messages, streaming, temperature, and max output controls.
API reference →Company
About LumeAPI
LumeAPI is an independent AI API gateway built for developers who need production access to GPT, Claude, Gemini, and other mainstream models — without maintaining a separate account, SDK, and billing relationship for every provider.
We operate a public portal at https://lumeapi.site for accounts, pricing, and documentation, and an OpenAI-compatible inference gateway at https://api.lumeapi.site/v1 for your applications. One key. One USD wallet. Text, image, and video.
38+Documented modelsText, image & video in one catalog
50%GPT savingsListed vs common official reference rates
40%Claude & GeminiOn supported catalog models
1API keyOne wallet across every modality
Our mission
AI inference has become infrastructure. Teams should be able to choose models, estimate cost, and ship features without re-architecting their codebase every time pricing or provider strategy changes.
LumeAPI exists to lower the operational friction of running AI in production: published rates you can plan around, a single integration surface your engineers already understand, and enough transparency in usage logs and wallet billing that finance and engineering stay aligned as traffic grows.
We focus on mainstream, production-grade models and clear documentation — not on listing every experimental checkpoint. That trade-off lets us optimize for reliable access, straightforward migration, and competitive catalog pricing on the models most products actually run.
What we offer
Everything below is available through the same account. You do not need a different key for Claude versus GPT versus an image model — only the model ID in your request changes.
Image generation
Sync image endpoints for GPT Image, Seedream, Grok image models, and related APIs with published per-call pricing in the catalog.
Browse image models →Video jobs
Async video workflows: submit a job, poll status, download MP4 when complete — Kling, Wan, Vidu, Veo, and other catalog entries.
Video integration →Transparent wallet
Top up with USDT on supported chains. Spend in USD across models. No opaque credits, no per-provider prepaid balances to reconcile.
Create account →Model research
Long-form guides on pricing, migration from OpenAI, OpenRouter, Claude, Gemini, and agent cost control — written for production teams.
Read research →Usage & keys
Create and rotate API keys, inspect per-call logs with tokens and cost, and monitor wallet balance from the portal console.
Open console →Portal vs gateway
Portal — lumeapi.site
Human-facing product for people on your team:
- Registration, login, and wallet top-ups
- API key management and balance overview
- Model catalog with official vs LumeAPI pricing
- Per-model integration docs and curl examples
- Research library and contact support
- Usage logs for auditing spend
Gateway — api.lumeapi.site
Machine-facing API your application calls:
POST /v1/chat/completionsfor text- Image generation and edit endpoints
- Async video submit, poll, and content download
- Bearer authentication with your LumeAPI key
- Optional SSE streaming on supported text models
- Exact model IDs documented per catalog entry
How teams get started
Create your account
Register on the portal with email. Your LumeAPI account is separate from other products and controls wallet access, keys, and usage history.
Fund the USD wallet
Top up with USDT on supported networks. Confirmed deposits credit your wallet in USD so you can spend across any catalog model.
Issue an API key
Generate a gateway key in Console. Keys are tied to your wallet balance and the models enabled for your account.
Point your app at the gateway
Set base URL to https://api.lumeapi.site/v1, pass your key as Bearer auth, and call the exact model ID from the catalog — same request shapes as the OpenAI SDK.
Ship and monitor
Track usage per call, adjust models as quality and cost require, and scale traffic with published rates instead of surprise invoices.
Who we build for
Product & SaaS teams
Ship chat, agents, content tools, or internal copilots without negotiating separate contracts for every model provider first.
Engineering leads
Standardize on the OpenAI SDK or compatible frameworks — including Cursor — and switch models by changing the model ID.
Cost-conscious builders
Compare official reference rates with LumeAPI catalog pricing before traffic scales, and route workloads to the right model tier.
Multimodal products
Combine text, image, and video generation behind one key and one billing view instead of three integration projects.
Pricing & economics
Every model in the catalog shows reference pricing alongside current LumeAPI rates. Supported GPT models are listed at roughly half of common official reference input/output prices; many Claude and Gemini entries are listed at a 40% reduction on standard rates shown in our public tables.
List price is only the starting point. Production AI cost depends on conversation length, agent loops, tool calls, retries, output size, and whether you need provider-native features such as Batch jobs or prompt caching. We publish research articles that model real monthly scenarios for chatbots, RAG, coding agents, and migrations from OpenAI or OpenRouter so you can estimate total spend — not just tokens.
The metric we recommend is cost per successfully completed task. A cheaper per-token rate that requires more retries or longer outputs may cost more in practice than a slightly higher rate with better completion quality.
Integration & compatibility
LumeAPI is designed as a drop-in layer for applications that already speak OpenAI's Chat Completions format. In most cases you change three values: API key, base URL, and model ID. Python, Node, curl, and tools like Cursor can target https://api.lumeapi.site/v1 without rewriting your prompt templates.
"OpenAI-compatible" does not mean every provider-specific feature is available. Hosted OpenAI tools, Anthropic Batch, Google Search Grounding, Realtime audio, and some caching controls may still require the official platform or additional engineering on your side. Our docs describe what each catalog model supports; test streaming, tools, and structured output before moving production traffic.
Independence & responsibility
- LumeAPI is not OpenAI, Anthropic, Google, or any upstream model company. We are a third-party gateway that routes requests to providers listed in our catalog.
- Model outputs are generated by upstream systems. You are responsible for how results are presented, stored, and used in your product, including safety, compliance, and intellectual property.
- Prompts and payloads transit our gateway to upstream providers. Review our Privacy Policy and Terms of Service before sending regulated or highly confidential data.
- We aim for high availability but do not guarantee uninterrupted service. Upstream latency, availability, and output quality can change without notice.
Ready to build?
Create an account, fund your wallet, and run your first request in minutes. Questions about billing, integration, or model choice? Our support team is available on the contact page.