Interactive API Explorer

The public developer schema holds the integrator's HTTP surface, 2 operations (POST /v1/chat/completions and GET /v1/models), browsable and runnable in your browser and generated directly from the live OpenAPI spec, so it follows the deployed schema.

Authenticating

Every operation takes an organization key (sk-org-…) and names the scope it needs in its security (OrgApiKey: ["chat:write"]). A signed-in user's token is answered 401 AUTH_TOKEN_INVALID. See Authentication:

bash
curl https://samreshuuu.ru/v1/models \ -H "Authorization: Bearer $SAMRESHUUU_API_KEY"

Requests sent from the explorer hit the production API with your key and run real turns against your quota. A missing or invalid key returns 401; a key without chat:write gets 403 AUTH_PERMISSION_DENIED. See Errors & rate limits for the full status-code and error-code contract.

Generate a Client

The raw spec is standard OpenAPI 3: feed it to any generator to scaffold a typed client.

bash
npx @openapitools/openapi-generator-cli generate \ -i https://samreshuuu.ru/api/v1/developers/openapi.json \ -g typescript-fetch \ -o ./samreshuuu-client

The same spec works with any OpenAPI-aware toolchain: swap the -g target for python, go, java, and so on, or point tools like openapi-typescript at the URL directly.

What's in the public surface

The spec covers the OpenAI-compatible API only. Connected services and code execution are not exposed over REST: reach them through an agent in chat, or directly with the tools of the MCP server. The native /api/v1 of the Samreshuuu apps is not part of the public surface.

Rate limits at a glance

Organization API keys have per-key, plan-based limits:

PlanRequests per minute, per key
Pro180
Max360

Limits may change; the pricing page has the current values. Rate-limit errors carry Retry-After when a delay is available; the API sends no X-RateLimit-* headers. A spent weekly limit answers 402 and is not retried. Full retry guidance is in Errors & rate limits.

Frequently asked questions

What does the interactive API explorer cover?

The integrator's HTTP surface: POST /v1/chat/completions and GET /v1/models, browsable and runnable in the browser. It is generated from the live developer OpenAPI spec, so it follows the deployed schema. Connected services and code execution are reached over MCP, not REST.

How do I authenticate a request?

Send Authorization: Bearer sk-org-... Each operation in the spec names the scope the key needs; both ask chat:write, and a key without it gets 403 AUTH_PERMISSION_DENIED.

Can I generate a typed client from the API?

Yes. The spec at /api/v1/developers/openapi.json is standard OpenAPI 3: feed it to any generator (openapi-generator, openapi-typescript, and so on) to scaffold a typed client in your language.

Are there rate limits on the API?

Yes. Organization keys are limited per key by the organization's plan; the table is on this page. Rate-limit errors include Retry-After when a delay is available. See Errors & rate limits for the full contract.

Was this page helpful?