Connect your tools
Use familiar Responses and Messages APIs with selected models. Provider credentials remain private.
Cooma TokenFactory
AI model access.
Business control.
Connect your teams and applications to selected AI models, with permissions, usage limits and spend visibility in one place.
Request a demo Customer sign in
A platform for your teams
Make access simpler.
Keep responsibility clear.
Choose who can use what
Set model permissions, daily and monthly limits, and expiry dates for each project key.
Understand consumption
Review usage and recorded charges by key, so you can see how your applications contribute to spending.
Security & control
Control access.
Understand the data path.
Project-key permissions and expiry help define access to models. Data handling also depends on the model service your application uses.
Confirm where requests are processed, whether prompts and responses are retained or used for training, and which provider terms apply before sending sensitive information. Access controls and usage records do not, by themselves, establish those protections.
Discuss model security requirementsFor developers
A familiar way
to connect.
Use your Cooma key and an approved model. TokenFactory handles access checks and usage tracking.
Model availability, pricing and deployment arrangements are confirmed for each customer. Platform imagery shows a demo environment.
from openai import OpenAI
client = OpenAI(
api_key="YOUR_COOMA_KEY",
base_url="https://api.tokenfactory.cooma.ai/v1",
)
response = client.responses.create(
model="<approved-model>",
input="Summarise this project brief.",
)Explore the architecture
Go inside
TokenFactory.
Explore an interactive illustration of how GPU infrastructure, inference and model access fit together.
The experience illustrates an architecture. Current services and deployment options are confirmed separately.
Launch the 3D experienceBefore we talk
Common questions.
What is Cooma TokenFactory?
TokenFactory provides access to selected AI models for teams and applications, with project-key permissions, usage limits, expiry dates and spend visibility. Model availability, pricing and deployment arrangements are confirmed for each customer.
Do I need to rent a GPU server to use TokenFactory?
TokenFactory is a separate model-access offering. Start with the models and applications you want to use; dedicated GPU capacity is a separate conversation. Explore GPU Cloud.
Which APIs and controls can my team use?
Applications can connect through Responses and Messages APIs using a Cooma project key and an approved model. You can set model permissions, daily and monthly limits, and key expiry, then review usage and recorded charges by key.
How are prompts and responses handled?
Data handling depends on the model service selected. Confirm processing locations, retention, training-use terms and provider conditions before sending sensitive information. Project-key controls do not by themselves establish those protections.
Start a conversation
Put intelligence within reach.
Tell us which tools, teams and applications you want to connect.