# Model discovery

> Documentation Index / 文档总索引  
> https://x2v.light-ai.top/llms.txt  
> Source: https://x2v.light-ai.top/en/api-docs#model-discovery  
> Locale: en  
> This page is optimized for Agent/LLM reading.

Discover capabilities at runtime instead of guessing model names or mutable limits.

## Task matrix

| Task | Purpose |
|---|---|
| `t2i` | text to image |
| `i2i` | image editing / image to image |
| `t2v` | text to video |
| `i2v` | image to video |
| `flf2v` | first/last frame to video |
| `t2av` | text to audio-video |
| `i2av` | image to audio-video |
| `ref2av` | reference media to audio-video |
| `s2v` | speech/audio-driven video |
| `animate` | character animation |
| `isr` | image super resolution |
| `vsr` | video super resolution |

## Discovery sequence

Call `GET /api/v1/model/list` for the catalog, then `GET /api/v1/model/specs?task=TASK` for the active submission contract. Add `model_cls=...` only after selecting a model. The specs response is ordered by the active pipeline and remains an array even for one selected model.

## Source priority

Use live model specs for models, media requirements, conditional parameters, dimensions, duration, and file limits. Use `/openapi.json` for field and response shapes. Use these Markdown pages for workflows and recovery guidance. Never rely on remembered constraints when a live spec is available.

## Related topics

[quick-start](/en/api-docs/quick-start.md) · [authentication](/en/api-docs/authentication.md) · [task-api](/en/api-docs/task-api.md) · [workflow-api](/en/api-docs/workflow-api.md) · [errors](/en/api-docs/errors.md)

OpenAPI: https://x2v.light-ai.top/openapi.json · Live models: https://x2v.light-ai.top/api/v1/model/list · Skill: https://github.com/ModelTC/LightX2V-Skills
