Skip to content
AnyRouter
Bring your own keys

Your keys. Every provider. One API.

Add your own provider keys and subscriptions (OpenRouter, Gemini, Groq, NVIDIA NIM, Z.AI and more) and call them all through one OpenAI-compatible URL. No markup on your keys, your own rate limits, automatic pooling and failover. Add a key in seconds, right here.

Top BYOK providers

Create a key on the provider's platform, then add it here. Free-tier providers first.

Step-by-step setup guides
Google AI Studio
Gemini models with a generous free tier
Free tier
11
OpenRouter
Hundreds of models, including free ones
Free tier
131
Groq
Very fast inference, free tier to start
Free tier
1
NVIDIA NIM
NIM endpoints with free starter credits
Free tier
11
Cerebras
Ultra-fast inference, free daily quota
Free tier
1
Mistral
La Plateforme with a free experimentation tier
Free tier
2
Ollama Cloud
Run open models on Ollama Cloud
Free tier
11
Z.ai
GLM models direct from Z.AI (Zhipu)
Paid
8
DeepSeek
Low-cost reasoning + chat models
Paid
3
OpenAI
GPT models on your own OpenAI account
Paid
27
Anthropic
Claude models on your own Anthropic account
Paid
13
xAI
Grok models direct from xAI
Paid
8

Models you unlock with your key

Add a provider key and these models become callable through your AnyRouter endpoint, with no extra wiring.

Google AI Studio
11 models
gemini-2.5-flash-litegemini-2.5-flashgemini-2.5-progemini-3-flashgemini-3.1-flash-lite+6
OpenRouter
131 models
claude-fable-5.1claude-fable-5claude-haiku-4.5claude-opus-4.1claude-opus-4.5+126
Groq
1 models
gpt-oss-safeguard-20b
NVIDIA NIM
11 models
deepseek-v4.1-flashmuse-glimmer-30bkimi-k2.6kimi-k3nemotron-3-embed-1b+6
Cerebras
1 models
gpt-oss-120b
Mistral
2 models
mistral-medium-3.5-128bmistral-small-3.1-24b-instruct

Bring a key from any of these

50 providers wired up and ready. Add a key and route through one API.

AI/ML API
AIHubMix
Anthropic
Antigravity (Google)
AutoJev
Baseten
Cerebras
Chutes
Cline
Cohere
CommandCode
DeepInfra
DeepSeek
Experiential
Fastino
Fireworks
GMICloud
Google AI Studio
Google Vertex
Groq
Hugging Face
LLM Gateway
LongCat
Mistral
Moonshot AI
MuleRouter
Nous Research
NovitaAI
NVIDIA NIM
Ollama Cloud
OpenAI
OpenCode Zen
OpenCode Zen Go
OpenRouter
OrcaRouter
OVHcloud
Poolside
QwenCloud
Sakana AI
SambaNova
SiliconFlow
StepFun
StepFun Step Plan
TypeSafe
Venice
Vercel AI Gateway
xAI
Z.ai
Z.ai (Coding Plan)
ZenMux

Donate idle keys, earn credits?

Got a key with spare rate limit? Lend it to the pool and turn idle quota into credits you can spend on any model.

Rolling out
Reward rate
5–8%
of notional cost, back as credits
Active donors
·
keys in the shared pool
Credits rewarded
·
paid to donors so far
1Add a BYOK key

Add a provider key like any other BYOK key. All 51 supported providers can be donated.

2Toggle 'Donate to pool'

Opt in with explicit consent and an optional rate cap. Opt out any time.

3Earn 5–8% back

When the pool uses your key, you earn credits, spendable on any model.

Estimate your earnings(5–8% of pool usage, back as credits)
$100/mo
+$5–$8
$250/mo
+$13–$20
$500/mo
+$25–$40
$1,000/mo
+$50–$80
up to / month
Example

Donate your free NVIDIA NIM key and your idle Z.AI coding-plan subscription. If the pool routes $200 of usage through them in a month, you earn $10–$16 back as credits, depending on your plan (5% on Go, 8% on Pro/Max). Heavier keys earn proportionally more, and you keep first call on your own key.

Opt-in per key, explicit consent, opt out anytime

How BYOK works

Encrypted & write-only

Keys are encrypted at rest with a per-row salt, validated live when you add them, and never returned or logged. You see only a masked preview.

Pooled & load-balanced

Add several keys per provider and choose a strategy: fallback, round-robin, weighted, or random. Burned keys are quarantined automatically.

Clear routing precedence

Per-request override → always-use keys → your other BYOK keys → managed capacity. You're billed by the provider; AnyRouter never charges credits for BYOK.

Full BYOK documentationPer-provider key guides