Provider accounts
Anthropic
Max 20×
OpenAI
Pro
01
Connect your plans
Add Claude and ChatGPT accounts to one encrypted provider pool.
AI subscription pooling
Pool your AI subscriptions. Share them with your team.
Connect Claude, ChatGPT, and other plans to one inference endpoint. Tokenmaxer balances usage so your team gets more from the capacity you already pay for.
- AI subscriptions pooled
- 100+
- in AI spend saved
- $50k+
Build your own Max plan
Build the plan your team actually needs.
Combine plans across accounts and providers. Add capacity based on usage, not headcount.
Build your pool
Claude Max 20×
Anthropic
Claude Max 5×
Anthropic
ChatGPT Pro 20×
OpenAI
ChatGPT Pro 5×
OpenAI
Your team capacity
Plans
4
Providers
2
Inference capacity
65×
Anthropic3 plans
Max 20××2Max 5××1
OpenAI1 Pro plan
Pro 20××1
One endpoint, shared across your whole team
EngineeringMarketingProductSales
How it works
Three steps. No workflow migration.
Setup takes less than 60 seconds.
Connected
Connected
Tokenmaxer keys
Engineering
sk-••••••••7f2a
CI and agents
sk-••••••••c91d
02
Create team keys
Issue scoped, revocable keys for people, teams, projects, or clients.
Tokenmaxer
Claude CodeCodex
03
Keep your tools
Point your existing harness at Tokenmaxer. Routing and failover happen behind one endpoint.
Capacity
Get more from your inference capacity.
Tokenmaxer routes traffic across your providers to maximize available capacity.
Provider capacity
Dashed markers show scheduled quota resets.
Range5h7d
Anthropic
90% pooled capacity left
- WorkMax 20×
- TeamMax 5×
- PersonalPro
Full capacity by Thu 6:41 PM
OpenAI
58% pooled capacity left
- TeamPro
- PersonalPlus
Full capacity by Sat 6:41 AM
Capacity-aware load balancing
Route requests against the available headroom and limit windows of each plan.
Higher effective limits
Every connected plan expands the pool's effective usage ceiling.
Automatic failover
Move traffic away from rate-limited or exhausted plans automatically.
Metered overflow
Fall back to pay-as-you-go inference when the entire pool is exhausted.
Scoped budgets
Set usage limits for each person, team, project, or client.
Usage telemetry
Attribute consumption, detect anomalies, and find capacity likely to expire unused.
Built for uneven demand
Capacity goes where the work is.
Tokenmaxer routes requests through plans with room across teams, time zones, and workloads.
Busy weeks
When one team's usage spikes, Tokenmaxer routes requests through available plans across the pool.
Teams across time zones
As different teams start work, they can draw from the same shared capacity.
Client demand spikes
Scoped keys and budgets let one client use available headroom without losing control.
FAQ
Is this allowed under my provider's terms?
Tokenmaxer routes your organization's traffic through subscriptions it owns; it does not supply or resell accounts. Provider terms vary, so review your plan before connecting it.
What happens when a plan hits its limit?
Tokenmaxer routes to another eligible plan, then uses metered overflow if the pool is exhausted.
Does Tokenmaxer store my prompts?
No. Tokenmaxer does not store prompts or responses. Your data stays between you and your provider. Tokenmaxer records only usage metadata for routing and accounting.
Which providers work today?
Claude and ChatGPT are live. GLM, Gemini, Grok, Kimi are next. The current list is on the providers page.
Put the plans you already own to work.
Connect your first plan in less than 60 seconds. Free for 7 days.