Getting started
From your first enquiry to your first agreement.
Start with the outcome
Megatron Wholesale helps organizations purchase actual AI inference services. You can compare models, request supplier quotes, purchase immediate capacity, or agree a future delivery window. Your agreement connects the model and supplier terms to the usage your team consumes.
If you are exploring a requirement, use Contact Us on the public website. You do not need an account to send an enquiry. A successful submission returns a reference after your message has been saved.
Your first five steps
These steps describe the buyer journey. Approved model providers follow Provider onboarding and Publishing capacity for supplier-only workflows.
- Create your account. Open the workspace and select Create account. Enter your name, work email, organization, password, and requested role.
- Compare models. Open Model catalogue. Search for a model or training house and check input/output prices, context limits, modalities, and source dates.
- Define delivery. Decide whether you need immediate capacity or a future start. Choose an explicit end time and a quantity of at least 100 MT — 100 million input-equivalent tokens.
- Request a quote. Open Requests for quote and create a request with the model, volume, dates, and any region, retention, or throughput requirements.
- Review the terms. A provider quote or firm offer must have available capacity. Commercial approval is required before acceptance. Once an agreement is active, your team can use scoped API access and follow consumption in Usage & reconciliation.
Choose the right role
Use Buyer to procure inference and Provider to supply approved endpoints and capacity. Broker / administrator work manages participants, supplier arrangements, and platform operations. Approved brokers have administrator-equivalent management authority. Public broker requests remain pending approval; purchasing for a client still needs the appropriate recorded buyer mandate.
On this isolated local workspace, Sign in keeps a Demo account dropdown with Buyer demo, Provider demo, and Broker / admin demo. Choose one and select Sign in. The server checks the private stored password and the approved role. Those passwords are not shown on the page. Use Use my own account to type an email and password. This demonstration access stays available here; it is not a public internet login.
See Account and roles for the approval and authority rules.
What to prepare
| Requirement | A useful starting point |
|---|---|
| Model | The exact model and version, or a shortlist to discuss |
| Workload | What your application needs the model to do |
| Volume | Expected input/output mix and total input-equivalent allowance |
| Timing | Start, end, and any peak periods |
| Delivery constraints | Region, retention, throughput, and required capabilities |
These details help providers quote the service you need. A large allowance alone does not guarantee a particular speed or concurrency.
Understand what you are seeing
The public catalogue contains reference information, including source prices. A model listing does not mean there is an executable supplier offer, a configured endpoint, or reserved capacity.
Your workspace begins with your organization’s real records. Empty lists mean that no visible records have been created yet. There are no sample balances or capacity commitments presented as live business.
Read your workspace overview
The signed-in Overview page greets you by name with summary cards: Agreements, Future deliveries, Open RFQs, and Recent API requests. Each card counts a recent window — the latest 500 visible agreements or requests and the latest 100 metering records — not your lifetime totals.
Below the cards, shortcuts matched to your role lead to your next task, and Recent agreements shows your five latest agreements with View all for the complete list.
Where to go next
Read Account and roles to understand access, Comparing models to research supply, and Understanding token prices before choosing an allowance.