Scope your useful workflow
Bring five examples, your brand, server and acceptance checks. The target AInferen API will be verified separately.
YOUR BRAND · PLANNED INFERENCE THROUGH AINFEREN
Your AI router. Your brand. Your service margin.
Self-deploy or agree a setup: child keys, lifetime usage budgets and one model procurement endpoint through the separate AInferen service (target verification pending). You sell your client a useful service at your agreed price.
From a useful idea to paid work
Choose a workflow and package
Agree the brand, one useful workflow, server, budget, timeline and acceptance checks. Starter/Launch/Growth are quote-required proposals.
Sell a useful result
Set your client price and keep your service margin under agreed partner terms. All model requests go through the separate AInferen service (target verification pending); actual costs and repeat use determine sustainability.
Choose the launch you need
Draft setup budgets. Scope, license, timeline, model availability and acceptance are agreed in a quote before payment. Inference and hosting are separate.
Starter · $200
Self-deployable router code under your brand: child keys, lifetime usage budgets and requests through the separate AInferen service (target verification pending). Delivery includes the archive, instructions and one agreed smoke test.
Launch · $500
Starter plus setup on one server you control, your name and visual branding, and one agreed integration: a Telegram bot or an OpenAI-compatible client. Acceptance follows an agreed test workflow.
Growth · $1,000
Launch plus one project listing on our site, three materials for voluntary sharing and source measurement: visit → test → brief → agreed order. One report after 14 days; third-party outreach and advertising are outside the scope.
This buys a scoped service setup. The archive is a self-deployable template; it does not process your customers’ money. Child keys and usage budgets are not separate payment accounts.
Larger purchase, lower proposed model multiplier
This is a draft ladder for each individual API top-up, not a live tariff. Purchases are not accumulated to reach a tier. A comparable official model price anchors the retail quote. Procurement cost and sufficient margin are verified separately; this makes no actual-cost claim. Paid inference stays closed until the supplier is verified.
| One API top-up, USD | Model-price share of reference | Status |
|---|---|---|
| $10–<$20 | 0.50 | Проект / согласование |
| $20–<$100 | 0.40 | Проект / согласование |
| $100–<$300 | 0.30 | Проект / согласование |
| $300–<$500 | 0.25 | Проект / согласование |
| $500+ | 0.20 | Проект / согласование |
Exactly $20 enters 0.40, $100 enters 0.30, $300 enters 0.25 and $500 enters 0.20. A multiplier is the share of the agreed reference for that model. It is not a setup discount or bonus credit. A quote states input/output prices per 1M tokens, the price source, limits and expiry. Future increases apply to new quotes; accepted terms are preserved.
Choose the model transparently
Exact model
Paid Exact model mode uses the named, verified model; if unavailable, the request fails rather than silently using a cheaper one. The current template forwards the chosen identifier and rejects a different identifier in the response. Verification of a specific paid supplier is still pending.
Smart Router · planned
Selection of another model requires separate explicit customer opt-in and disclosure of the actual model used. Smart Router is not implemented in the current template.
Cache · planned
Cache policy must be visible beside the mode. Sensitive prompts and context-mismatched prompts are excluded; a cached result must not be presented as a new premium-model response. Caching is not implemented in the current template.
Your brand. Your client price.
Create child keys, set lifetime usage budgets and procure all inference through the separate AInferen service (target verification pending). Set your client price and keep your service margin under agreed partner terms. License and partner terms are agreed; this is not an automatic revenue-sharing system.
Source kit-v1 uses the historical HTTPS alias api.hesoyam.business; its fixed endpoint has not been silently replaced. API migration and alias availability are checked separately. The archive is immutable; it is not a deployed customer service.
The target API is verified separately
The last verified hs-mini demo used local Qwen2.5-0.5B q4_0 for brief text checks with weak quality. This does not promise availability at the new domain or paid-model quality.
5 attempts per key per rolling 24 hours, with shared network and global caps. Read the API terms for all request and registration limits.
Live limits and availabilityBefore the first step
What can I sell to my client?
An agreed service under your brand: a useful workflow, setup and support. Router license and your margin are agreed in the terms; paid model supply is not yet verified.
Do I need email or Telegram for the API?
No. An agent or human privately generates and saves a key, then registers api_key. A label is optional. Never put the key in URLs, public posts or prompts.
Are the 0.50–0.20 multipliers live?
No. They are a proposal for quotes after verifying cost and a comparable reference for the specific model. Consult API models and payment terms for runtime availability; there is no active paid tariff.
Verify supply, then pay
Paid inference and new crypto invoices are closed today. Checkout opens after model, price, supplier and payment readiness are verified. BTC/LTC are prepared; receiving instructions belong only to a live invoice. This proposal has no recurring top-up processing surcharge.
Bring a useful outcome to a real buyer
Show your operator a useful workflow, then agree a package. A real paid and accepted delivery is a first demand signal. Margin after actual costs can fund capacity; registrations are not customers.
Plan a partner projectTell us what you want to build
Copy the request and send it to support. Agree the useful outcome, model availability, scope and acceptance before payment.
I would like a quote for a HESOYAM router package. Package: Starter $200 / Launch $500 / Growth $1,000 My client outcome and one required integration: … Server and brand: … Separate inference budget / volume: … Timeline and acceptance checks: … Source: web_router
The link opens @HESOYAM_AI_SUP_BOT. You send the request yourself.