DeepSeek HarnessOne key. Your agent is ready.
Stop rebuilding provider plumbing for every agent. Connect DeepSeek V4 Flash Vision, Flash, and Pro through one endpoint, with image understanding, files, thinking, tools, and streaming.
3
V4 models
1M
Context window
1
Unified key
$ npx @deepseek-ai/dsh web
provider: ok-api
baseURL: https://api.ok-api.net/v1
model: deepseek-v4-flash-vision-exp
thinking: max
tools: enabled
✓ Agent session readyDeepSeek Harness is currently a Developer Preview and may introduce breaking changes. The OK-API integration guide will track official releases.
Why OK-API
More than a DeepSeek endpoint
Use the official DeepSeek API when a single provider is all you need. Choose OK-API when your agent needs one integration across API formats, operational visibility, and provider fallback.
One integration
OpenAI, Anthropic, and Responses-compatible request formats.
Operational controls
Per-key quotas, model restrictions, usage history, and spend visibility.
Resilient routing
Retry and failover across configured channels when an upstream is unavailable.
Transparent policies
Published billing, data handling, model routing, refunds, and service health.
1 Key
Unified authentication
3 Models
Vision / Flash / Pro
3 Formats
OpenAI / Anthropic / Responses
1M
Token context
Everything you need to ship AI
A comprehensive gateway that handles the complexity of multi-provider LLM integration.
Multi-Provider Gateway
Route requests to any major LLM provider through a single OpenAI-compatible endpoint.
Automatic Format Conversion
Use OpenAI, Claude, or Gemini format interchangeably — we convert and adapt automatically.
Smart Load Balancing
Weighted random routing, automatic failover, and configurable retry policies for high availability.
Usage Analytics
Track token consumption, request volume, and spending with real-time dashboard and charts.
Access Control
Per-token quotas, model restrictions, expiration dates, and role-based admin management.
Streaming & Realtime
Full SSE streaming support with automatic retry and error recovery for production workloads.
Supported Models & Providers
Seamlessly connect to every major LLM provider through a single gateway.
OpenAI
Anthropic Claude
Google Gemini
DeepSeek
ByteDance Seedance
Azure OpenAI
Meta Llama
Mistral AI
Cohere
Jina
Groq
Together AI
OpenRouter
Perplexity
xAI Grok
GitHub ModelsCustomPay only for the Credits you use
No recurring fee today. Top up at 1 USD = 1 Credit, then use Credits across available API models. Membership is coming soon.
Membership plans
Coming soon
Membership pricing and benefits are being finalized.
Ready to simplify your AI infrastructure?
Join developers and teams who use your-api to manage their LLM API calls.
Get Started FreeFrequently asked questions
What is your-api?+
your-api is a high-performance AI gateway that aggregates and routes requests to multiple LLM providers (OpenAI, Claude, Gemini, DeepSeek, and more) through a single unified API endpoint.
Which models and providers are supported?+
We support current OpenAI, Anthropic, Google, and DeepSeek models, including DeepSeek V4 Flash Vision, plus Seedance 2.0 video generation. Check the live model catalog for availability.
How does format conversion work?+
your-api automatically converts between OpenAI, Claude, and Gemini API formats. You can use any SDK with your preferred format, and we handle the translation to the upstream provider.
How is pricing calculated?+
Requests use prepaid Credits at the rate for the selected model. Text models use token-based or per-request pricing; Seedance video tasks use generated-video token pricing. Check the model catalog and promotion details before calling the API, and set API-key quota limits to control spending.
Which API key should I use?+
Create an OK-API key in your account and use it with the OK-API endpoint. Customers do not need to supply upstream provider keys. Upstream channel configuration is managed by platform administrators.
Do I need a membership?+
Membership plans are coming soon. For now, add prepaid Credits whenever you need them at 1 USD per Credit, then use that balance for API requests.
Contact
- support@your-api.xyz