OpenAI-compatible · 22 models · 8 providers
One API for every major LLM.
Linka is a unified inference gateway. Route to OpenAI, Anthropic, Google, DeepSeek, Meta, Mistral, Qwen, and xAI through a single OpenAI-compatible endpoint — prepaid credits, per-token pricing, no subscriptions.
curl "https://api.linka.ink/v1/chat/completions" \
-H "Content-Type: application/json" \
-H "Authorization: Bearer sk-linka-your-key" \
-d '{
"model": "openai/gpt-4o",
"messages": [
{ "role": "system", "content": "You are a concise assistant." },
{ "role": "user", "content": "Explain an inference gateway in one sentence." }
]
}'
# Response (truncated)
# {
# "id": "chatcmpl-linka-9f2c71",
# "model": "openai/gpt-4o",
# "choices": [ { "message": { "role": "assistant", "content": "An inference gateway ..." } } ],
# "usage": { "prompt_tokens": 41, "completion_tokens": 28, "cost_credits": 0.00038 }
# }Why Linka
The shortest path between your code and every frontier model.
One API, every model
Switch models by changing a single string. Linka normalizes request and response shapes across providers, so your integration never forks per vendor.
Prepaid credits, per-token pricing
Top up any amount and spend it down at published per-model rates. 1 credit = $1 of usage, deducted in real time. Credits never expire.
OpenAI-compatible drop-in
Point the official OpenAI SDK at api.linka.ink/v1 and keep your code. Works with the Python and Node SDKs, LangChain, LlamaIndex, and anything that speaks /chat/completions.
Smart routing & fallbacks
Pass a fallback chain and Linka reroutes automatically when an upstream provider is down, slow, or rate-limited. You are billed at the rate of the model that served the request.
Usage analytics
Per-key, per-model token and spend analytics in the console. Set per-key spend caps, export CSV, and query your balance programmatically.
No subscriptions, ever
No monthly minimums, no seat fees, no tiers to unlock. You pay only for the tokens you use, from the balance you already funded.
Supported models
Frontier, open-weight, and everything between.
Every model is available on the same endpoint, billed per token from the same prepaid balance. Prices in USD per 1M tokens.
| Model | Context | Input / 1M | Output / 1M |
|---|---|---|---|
| openai/gpt-4oPopular | 128K | $2.50 | $10.00 |
| anthropic/claude-sonnet-4.5Popular | 200K | $3.00 | $15.00 |
| google/gemini-2.5-proPopular | 1M | $1.25 | $10.00 |
| openai/gpt-4o-miniValue | 128K | $0.15 | $0.60 |
| google/gemini-2.5-flashValue | 1M | $0.30 | $2.50 |
| deepseek/deepseek-chat-v3Value | 164K | $0.27 | $1.10 |
| deepseek/deepseek-r1Reasoning | 164K | $0.55 | $2.19 |
| meta-llama/llama-3.3-70b-instructValue | 131K | $0.12 | $0.30 |
| anthropic/claude-opus-4.1 | 200K | $15.00 | $75.00 |
| x-ai/grok-3 | 131K | $3.00 | $15.00 |
Pricing
Prepaid credits. Per-token rates. Nothing else.
- 1 credit = $1 of usage — top up any amount from $10
- Per-token rates deducted in real time at published per-model prices
- Credits never expire
- Volume bonuses: +2% on top-ups of $500+, +5% on $2,000+
- Unused credits refundable within 14 days of top-up
Worked example
Your app sends 2,000 requests/day, averaging 600 input and 250 output tokens on openai/gpt-4o-mini:
FAQ
Questions, answered.
base_url (Python) or baseURL (Node) to https://api.linka.ink/v1, use your sk-linka-... key, and the official SDKs work unchanged — including streaming. See the docs for examples.Route your first request in under two minutes.
Create an account, top up any amount, and call any model on the same endpoint. Credits land instantly after payment.