PortageAI

Usage-based AI API · Billed monthly in CAD

Pay per token. Never per seat.

PortageAI gives your whole team access to leading AI models through one OpenAI-compatible API. There are no seat licences and no minimum commitment. Each month you get one invoice for exactly what you used.

Request access How billing works

Invite-only while we onboard our first teams.

October usage Example figures
ModelTokensCAD
Fast chat model3,420,000$4.87
Reasoning model610,000$9.12
Embeddings12,800,000$0.91
Seats (14 people)—$0.00
Total16,830,000$14.90
One line per model. Fourteen people used it this month; nobody needed a licence.

How billing works

From request to invoice

  1. Every request is metered

    Each call records the model, its input and output tokens, and its cost. Nothing is estimated or rounded up to a plan tier.

  2. Usage rolls up by team

    Give each team, app or client its own key. Usage and spend stay separate, so you can see exactly where it goes.

  3. One invoice a month, in CAD

    At month end you get a single invoice with a line per model. There is no seat count to reconcile and no currency conversion on your side.

Why no seats

Your bill follows your usage, not your headcount

Seat licences

  • You pay for every person with a login, whether they use it daily or twice a quarter.
  • Adding a contractor or a new hire means buying another licence.
  • Unused seats are spend you never get back.

PortageAI

  • You pay for the tokens your requests use, and nothing else.
  • Connect as many people, apps and workflows as you like.
  • A quiet month costs less. A busy month is never a surprise.

Pricing

Usage only. That's the whole plan.

Every model has a per-token rate based on its provider's pricing. You get the full rate card before you send a single request, and your invoice shows each model on its own line.

Per token
Model usage
per-token rate card
Seat fees
none
Platform fee
none
Minimum commitment
none
Billing
monthly invoice, CAD
Request the rate card

Spend controls

Usage-based without the open tab

Paying per token only works if you stay in control of the total. Limits are set per key, so one busy project can't run up the whole bill.

  • Monthly budgets per key or team

    When a key reaches its budget, requests stop instead of running up charges. Raise it any time.

  • Model allow-lists

    Choose which models each key can call, from fast and inexpensive to the most capable.

  • Usage by model, key and day

    Check any key's spend through the API, or ask us for a breakdown by team or client before the invoice arrives.

  • Automatic fallbacks

    If a model provider has an outage, requests can move to a backup model so your app keeps working.

Developers

Change one line

PortageAI speaks the OpenAI API. Point your existing SDK at our endpoint and keep the rest of your code. It works the same from curl, LangChain, LlamaIndex and any tool that accepts an OpenAI-compatible base URL.

python
from openai import OpenAI

client = OpenAI(
    base_url="https://api.portageai.app/v1",  # the only change
    api_key="sk-your-portage-key",
)

reply = client.chat.completions.create(
    model="fast-chat",
    messages=[{"role": "user", "content": "Summarize this contract"}],
)
print(reply.usage.total_tokens)  # what you're billed for

Request access

Get a key for your team

Tell us a little about what you're building. We'll reply with your rate card and next steps.

We use these details only to respond to your request.