One API for 130+ AI models
OpenAI, Anthropic, Google, xAI, DeepSeek and more.
Stop juggling provider accounts, API keys and separate invoices. One key, one balance, one OpenAI-compatible format, across text, image, video and live web data.
- No credit card required
- Setup in 30 seconds
- No subscription
Models from
Context for AI
Your agents can read the web. Not just models.
Clean, LLM-ready output on the same key and the same balance as every model. Failed requests are not billed.
POST /v1/scrapeRead a page
Any URL to clean markdown, with automatic anti-bot handling.
Learn more →POST /v1/crawlRead a site
Crawl a whole website into markdown as an async job.
Learn more →POST /v1/extractStructured extract
Ready-made JSON from Amazon, Crunchbase, Google Maps and more.
Learn more →No other model gateway gives you this on the same key.
Beyond text
Image, video and audio through the same API
The same key, the same balance and the same real-dollar pricing as your chat models.
29 image models
Flux, Imagen, Seedream, Ideogram, Recraft and Stability, priced in real dollars per image.
Browse image models →26 video models
Veo, Kling, Seedance and Hailuo as async jobs, priced per second of output.
Browse video models →Speech and audio
ElevenLabs voices on the same key and the same balance as every other model.
Browse audio models →Drop-in compatible. Zero learning curve.
Already using the OpenAI SDK? Change one line, your base URL. Works with any language or framework.
from openai import OpenAI
client = OpenAI(
base_url="https://api.apexapi.dev/v1",
api_key="<YOUR_APEXAPI_KEY>",
)
response = client.chat.completions.create(
model="anthropic/claude-opus-5",
messages=[{"role": "user", "content": "Hello!"}],
)
print(response.choices[0].message.content)Using a coding agent? Cursor, Cline, Codex CLI, Copilot and 8 more guides.
Reliability
When a provider goes down, your app does not
A single provider outage is the most common way an AI feature breaks. Models that run on more than one rail fail over inside the same request, with no code change on your side.
We route through direct provider APIs plus compliant infrastructure rails, so models have somewhere to fall back to:
- Anthropic direct→Anthropic on Google Vertex
- Google AI Studio→Google Vertex
- Mistral direct→Mistral on Google Vertex
- Alibaba direct→Alibaba on Google Vertex
The published price of a model is the price of its most expensive rail, so a failover never costs you more than you were quoted.
Everything you need to integrate AI
A single API key unlocks every major AI model. No more managing multiple provider accounts.
Unified API
One OpenAI-compatible format for every model. Switch providers by changing a string, not your code.
Real dollars, not credits
$1 buys $1 of credits. No bonus tiers, no expiring balance, no invented currency to convert in your head.
Streaming everywhere
Full Server-Sent Events on every chat model, so tokens reach your users as they are generated.
Cost on every request
The exact price of a call comes back on the response and lands in your dashboard logs, streaming included.
Guardrails built in
PII detection, prompt-injection blocking, per-key rate limits and model allowlists, on by default.
Agent-native
A hosted MCP server, agent self-registration and autonomous USDC funding over x402. An agent can get its own key and top up its own balance.
Built for teams
When AI spend stops being a side project, you need control over who spends what.
- Organizations and roles
- Shared credit balance
- Per-member spend caps
- Per-member model access
- Per-key limits and allowlists
- Team-wide usage and request logs
Start building with AI today
Create an account, add credits, and make your first API call in under 2 minutes.
No credit card required. No subscription. Credits never expire.