SDK glossary¶
Topics: architecture, generation, requests, workers
| Term | Meaning in Horde SDK |
|---|---|
| API model | A Pydantic object representing a request, response, or nested wire value. |
| Client session | A connection-owning client that submits individual requests and tracks follow-up cleanup. |
| Dispatch | Selection and conversion that turns popped Horde work into backend-ready parameters. |
| Dry run | An API request that returns estimated kudos without creating generation work. |
| Generation | One produced image, text, or alchemy result within a request or worker job. |
| Manual client | A client exposing individual submission, poll, status, and delete operations. |
| Parameter model | A reusable typed representation of generation intent, separate from a particular endpoint. |
| Simple client | A convenience client that owns polling, final retrieval, timeouts, and cleanup. |
| Worker job | The SDK object coordinating one popped request and its individual generations. |
The shared Horde definitions cover ecosystem terms such as worker, bridge, kudos, model, request, and job. When a shared and SDK-specific term overlap, the shared definition describes the system concept and this page describes its Python representation.
Code map¶
| Concept | Module | Symbol |
|---|---|---|
| Request | horde_sdk/generic_api/apimodels.py |
HordeRequest |
| Simple client | horde_sdk/ai_horde_api/ai_horde_clients.py |
AIHordeAPISimpleClient |
| Worker job | horde_sdk/worker/job_base.py |
HordeWorkerJob |