How do I get an API Key?

Register and sign in, then create an API Key from Keys in the console.

Which models and providers are supported?

OpenAI, Anthropic, Gemini and many compatible models are supported. Check Model Square for current availability.

How does billing work?

Usage is billed by model input, output, or request volume. See Model Square for current prices.

What are the request limits?

The default rate limit per user is 60 RPM; for special needs, contact live chat to have it raised. Usage and request logs are in the console. No SLA is offered to individual users; for business customers, service commitments follow the Enterprise Service Agreement.

Aggregated API Gateway
built for massive AI model coverage

Q: Can we sign a contract?

Yes. Please contact our business team to confirm the contracting party, contract terms and governing law. If you have special requirements, please get in touch.

Q: Do you use official APIs?

It depends on the group. Upstreams of public groups include official model-vendor APIs, cloud providers and third-party API providers. We also operate our own NVIDIA® B300 inference clusters (Inner Mongolia, Hong Kong and Taiwan, China; Los Angeles, USA). To confirm the upstream type of a specific group, ask sales for details.

Q: How is data handled? Is it used for training?

Requests are sent to the upstream of the group you choose for processing. We do not use request content to train models of our own. Usage records do not include prompts or response text; retention periods for each type of log are set out in the Privacy Policy. How upstreams handle data is governed by their own terms of service.

Q: Can rate limits be raised?

Yes. The default rate limit per user is 60 RPM; for special needs, contact us to have it raised, and dedicated groups are evaluated and enabled on request. Rate limits are gateway limits — actual throughput also depends on the upstream, so higher limits need load testing per model first.

Clear billing, stable routing, high-performance calls.

ReliableHigh availability, low latency Flexible routingIntelligent multi-model dispatch Developer friendlyComplete docs and examples

Get Started View Docs

200 OK

POST/v1/chat/completions

REQUEST

curl -X POST "/v1/chat/completions" \  -H "Authorization: Bearer sk-••••" \  -d '{    "model": "your-model",    "messages": [      { "role": "user", "content": "..." }    ]  }'

RESPONSE

{  "choices": [{ "message": { "content": "Chat request routed." }}],  "usage": { "total_tokens": 27 }}

Telegram community

Clients / IDE Integrations

Setup guides for 32 clients, including Claude Code, Codex, VS Code, Cline and Cherry Studio · CC Switch guide

Connect in 3 Steps

1 Create account Register and obtain an API Key

2 Choose a model Find a model in Model Square

3 Call the API Integrate it into your app or tool

View quick-start guide

Meet 88API

Run by a Hong Kong–registered company — an AI model gateway for developers and businesses worldwide.

Legal name: 88 DIGITAL TECHNOLOGY LIMITED
Registered in: Hong Kong SAR, China
Business Registration No.: 80862628
Company address: RM 18, 2/F, TusPark, 118 Wai Yip Street, Kwun Tong, Hong Kong
Own compute: NVIDIA® B300 inference clusters (Inner Mongolia, Hong Kong and Taiwan, China; Los Angeles, USA)

Verify with the Hong Kong Companies Registry Contact sales

Contracts, billing and cost control

Corporate billing, higher rate limits and cost control — all in one account.

Corporate billing and invoicing

Corporate billing is arranged through live chat, where our team will share payment details and the process; invoicing can be requested online in the console and is issued directly.

Rate limits and dedicated groups

Higher enterprise RPM limits and SLA-backed stability; dedicated enterprise groups set up on request.

Multi-key cost control

Create API keys per project or member, each with its own quota, allowed models, IP allowlist and expiry.

Enterprise

Contracts, corporate billing and per-project cost control.

Enterprise Service Agreement available; corporate billing via live chat; invoicing online in the console
Higher enterprise RPM limits and SLA-backed stability; dedicated enterprise groups set up on request
Multi-key cost control and usage reports

Explore Enterprise

Reseller Partners

Use 88API as the upstream for your API site, relay service or client systems.

OpenAI, Claude and Gemini compatible; connect by adding a channel
Partner-only pricing, RPM limit lifted, custom model parameters passed through in full
Integration plans assessed one-on-one by our business team based on volume

Explore Reseller Partners