# Kvantok

> AI API for products and teams with payments and balance in Russian rubles.

- **Canonical URL**: https://kvantok.ru/
- **Markdown twin**: https://kvantok.ru/index.md
- **Primary audience**: product teams, SaaS developers, bot builders, support teams, internal automation teams.
- **Primary use case**: buy access to modern AI models through one API and pay from a RUB balance.

## What Kvantok Does

Kvantok provides access to AI models through a unified API and a client workspace. Customers top up a ruble balance, create API tokens, choose public model IDs, set limits, and track usage without managing separate provider contracts or foreign-currency payments.

## Key Benefits

- One API for model access.
- Customer-facing balance, payments, usage history, and statements are RUB-only.
- API tokens can be created, monitored, limited, and revoked.
- RPS/RPM limits can be applied by customer, IP, and token.
- OpenAI-compatible and Anthropic-compatible request formats are supported.
- Admin/provider routing and supplier cost details are kept outside the client-facing workspace.

## Main Customer Scenarios

- **SaaS and AI features**: add AI responses, summaries, classifiers, and assistants to products.
- **Bots and support**: use fast and cost-efficient models for dialogs, ticket classification, and summaries.
- **Internal tools**: give developers and operations teams controlled model access for documents, analytics, coding workflows, and automation.

## Pricing Model

Customers see prices and balances in rubles. Public starting prices shown on the website include model-level input and output pricing per 1M tokens. The current customer catalog and final availability are shown in the workspace.

## Public Model Examples

- `gpt-4.1`: strong model for complex tasks.
- `gpt-4.1-mini`: fast and cost-efficient model for mass usage.
- `claude-sonnet-4`: reasoning-oriented model for text, analysis, and product tasks.
- `codex-pro`: coding model for development workflows.

## API Integration

Developers can start from the technical API guide. The public website documents only methods that are useful from a client API token.

Important endpoints:

- `GET /v1/models`: list enabled public models.
- `POST /v1/chat/completions`: OpenAI-compatible chat completions.
- `POST /v1/responses`: Responses-compatible endpoint.
- `POST /v1/messages`: Anthropic-compatible messages endpoint.

## Security And Finance Controls

- Login and API tokens are separated.
- 2FA is requested only after valid email and password when enabled.
- Raw prompts and raw model responses are not stored by default.
- Balance mutations are ledger-backed.
- Streaming requests reserve balance before provider calls.
- Payment webhooks are handled idempotently.

## FAQ

### Does Kvantok show USD prices to customers?

No. The customer workspace is RUB-only. Supplier currency and provider costs are internal accounting details.

### Can developers use OpenAI-compatible SDKs?

Yes. Developers can point compatible SDKs to the Kvantok base URL, use a public model ID, and authenticate with a client API token.

### How does Kvantok help control spending?

Customers can use separate API tokens, RPS/RPM limits, usage history, low-balance notifications, and RUB statements.

## Related Pages

**Documentation:** [API guide](https://kvantok.ru/docs/api) | [API-key methods](https://kvantok.ru/docs/public)
**See also:** [Pricing](https://kvantok.ru/pricing) | [Security](https://kvantok.ru/security) | [Create account](https://kvantok.ru/signup)