Experiential
13.2k stars on GitHubSign inStart for free

Your AI bill
is the plan price.

No per‑token charge on DeepSeek V4.1 Flash, GPT‑6 Luna and Jev. Pick a plan and the price stays the same every month.

Start for freeBook a call

Free

$0/mo

  • Unlimited DeepSeek V4.1 Flash
  • Unlimited GPT-6 Luna
  • Unlimited Jev
Start free

Up to 240 requests/min

Pro

$20/mo

  • Everything in Free
  • Your own provider keys
  • Zero data retention
Start on Pro

Up to 1,500 requests/min

Max

$200/mo

  • Everything in Pro
  • 75% off DeepSeek V4 Flash and Qwen3.8 27B
  • Early releases
Start on Max

Up to 6,000 requests/min

Ultra

$2,000/mo

  • Everything in Max
  • DeepSeek V4 Flash and Qwen3.8 27B at $0
  • 45% off Kimi K3
Start on Ultra

Up to 10,000 requests/min

Each plan sets your speed. Every plan runs every model.

PlanRequestsTokensOther models
Free$0/mo10/min250K/minNo creditsStart free
Pro$20/mo1,500/min25M/min$20 includedStart on Pro
Max$200/mo6,000/min100M/min$200 includedStart on Max
Ultra$2,000/mo10,000/min150M/min$2,000 includedStart on Ultra

On every plan DeepSeek V4.1 Flash, GPT-6 Luna and Jev have no per-token charge. Credits cover usage on other models at 1 cent per credit.

Free limits 10 shared requests/min, 250K shared tokens/min and 1M tokens/day across DeepSeek, Luna and Clef. No general credits.

Three models with no per-token charge. On every plan including Free.

On every plan

The gateway comes with the plan. Free included.

  • One key for 963 models
  • OpenAI-compatible with streaming
  • Fallback across providers in an order you set
  • Budgets that stop requests at the cap
  • Usage by key, person, model and day
  • Request logs with route and cost

A busy month costs the same. Your bill stays the plan price.

You know the number before the month starts. Usage of the three models never adds to it.

Questions

What does unlimited mean?
DeepSeek V4.1 Flash, GPT‑6 Luna and Jev carry no per-token charge on any plan. Your bill is the plan price however much you run them.
Is Free really free?
Yes. The three models cost nothing on Free. Free has shared per-minute and daily caps.
Is there a rate limit?
Each plan caps requests and tokens per minute. Free core models share 10 requests and 250K tokens a minute, with 1M combined tokens/day. Ultra allows 10,000 and 150M. Requests above a cap get a 429 response.
Can I bring my own provider keys?
Yes on Pro and up. Traffic on your own keys is billed by your provider and we never rate limit it.
Can I pay yearly?
Yes. Pay for 12 months up front.
What if I need more?
Talk to us about Enterprise: committed pricing for the whole company.

Other models

The other 960 models bill per token at the provider's price with nothing on top. Plans discount some of them. Paid plans include monthly credits and 1 credit is 1¢. Browse the models

Free
No general creditsOne-time $1 card check for free models
Pro
2,000 credits/mo
Max
20,000 credits/mo
Ultra
200,000 credits/mo
Discounts on other models by plan
ModelFreeProMaxUltra
DeepSeek V4 Flash25% off50% off75% off$0
Qwen3.8 27B25% off50% off75% off$0
Kimi K3List11.25% off22.5% off45% off
GLM 5.3List7.5% off15% off30% off
Claude Sonnet 5.5ListList20% off20% off
GPT-5 family and GPT-4.1List5% off10% off20% off
Qwen3.8 MaxList3.75% off7.5% off15% off
Claude Fable 5, Fable 5.1 and Sonnet 5List2.5% off5% off10% off
GLM-5.3 FlashList2.5% off5% off10% off
Qwen3.8 Flash and Qwen3.8 2.4T A95BList2.5% off5% off9.1% off
TTencent Hy4 PreviewList2.5% off5% off9.1% off
Claude Opus 4.5 to 4.8, Sonnet 4.5 and 4.6, Haiku 4.5List1.25% off2.5% off4.05% off

Each discount covers your plan's monthly credits of usage. Then list price. Every model page shows its price on each plan.

Committed pricing for the whole company

Our lowest rate with private networking, data residency and SSO.

EnterpriseBook a call

Start free. One flat price after that.

Start for freeBook a call