# Public API reference

Source: https://provod.ai/en/docs/reference

*Public API reference*

Use the API base URL with `/v1` and a Bearer key for the selected workspace. Parameters and support depend on the model. Inference keys and management keys are not interchangeable.

| Method and path               | Purpose            | Guide                                         |
| ----------------------------- | ------------------ | --------------------------------------------- |
| `GET /v1/models`              | Catalog            | [Models](/en/docs/catalog)                    |
| `GET /v1/models/count`        | Model count        | [Catalog](/en/docs/catalog)                   |
| `POST /v1/chat/completions`   | Chat and streaming | [Chat Completions](/en/docs/chat-completions) |
| `POST /v1/responses`          | Responses          | [Responses](/en/docs/responses)               |
| `POST /v1/messages`           | Messages           | [Messages](/en/docs/anthropic-messages)       |
| `GET /v1/images/models`       | Image models       | [Images](/en/docs/images)                     |
| `POST /v1/images/generations` | Image generation   | [Images](/en/docs/images)                     |
| `GET /v1/videos/models`       | Video models       | [Video](/en/docs/videos)                      |
| `POST /v1/videos`             | Create video job   | [Video](/en/docs/videos)                      |
| `GET /v1/videos/{id}`         | Video job state    | [Video](/en/docs/videos)                      |
| `GET /v1/videos/{id}/content` | Download video     | [Video](/en/docs/videos)                      |

See the Messages guide for its header requirements. This short reference does not describe the private cabinet or administrative API. Manage keys, balance and usage in the corresponding cabinet sections.

[Authentication](/en/docs/authentication) · [Errors](/en/docs/errors) · [Limits](/en/docs/limits)

## Troubleshooting


**The call returns 401 or 404**


Check the base URL, path, method and key type against the relevant guide. Do not use an agent token in place of an inference key.
