# Descript

> Connect Descript to automate video and audio workflows.

App · 14 actions. Web page: https://docs.lodol.com/docs/integrations/descript

## Actions

### Agent edit

Agent edit.

**Inputs**

- Project (`project_id`, text): The ID of an existing project to edit. Mutually exclusive with `project_name`.
- Project name (`project_name`, text): Name for creating a new project. Mutually exclusive with `project_id`.
- Composition id (`composition_id`, text): Composition to target within the project. When provided, the agent will focus its edits on this specific composition rather than choosing one automatically. Only valid when `project_id` is also provided. Requires `project_id`. Accepts any of the following formats: - A full composition UUID (e.g. `39677a40-1c43-4c36-8449-46cfbc4de2b5`) - A 5-character short ID from a Descript URL (e.g. `39677`) - A full Descript project URL (e.g. `https://web.descript.com/{project_id}/39677`)
- Model (`model`, text): AI model to use for editing. Accepts a canonical model id (e.g. `claude-opus-5.5`) or a friendly alias that tracks the stable version of a family (e.g. `claude-opus`). Call [GET /agent/models](#operation/listAgentModels) for the current set of supported models and aliases. Defaults to `auto` when omitted, which selects a recommended model for your account.
- Prompt (`prompt`, text, required): Natural language instruction for the agent to execute. Examples: "add studio sound to every clip", "remove all filler words", "create a 30-second highlight reel"
- Team access (`team_access`, text): Access level for team members when creating a new project. Only applicable when `project_name` is provided (not when using `project_id`). Defaults to `none` if not specified. One of: `edit`, `comment`, `view`, `none`.
- Callback url (`callback_url`, text): Optional webhook URL to call when the job completes or fails. Descript will POST the job status (same format as [GET /jobs/{job_id}](#operation/getJob)) to this URL.
- Conversation id (`conversation_id`, text): Conversation ID from a previous agent response. Pass it back on your next call, with the same `project_id`, so the agent sees prior turns. Omit it to start a new conversation. Requires `project_id`.

**Outputs**: Result returned by the service.

- `response` (object): Complete service response, including pagination.

**Example output**

```json
{
  "response": {
    "conversation_id": "example",
    "drive_id": "example",
    "drive_name": "example",
    "job_id": "example-job",
    "project_id": "example",
    "project_url": "example",
    "resolved_model": "example"
  }
}
```

### Cancel job

Cancel job.

**Inputs**

- Job id (`job_id`, text, required): The job ID

**Outputs**: Result returned by the service.

- `response` (object): Complete service response, including pagination.

**Example output**

```json
{
  "response": {}
}
```

### Check API status

Check API status.

**Inputs**: None.

**Outputs**: Result returned by the service.

- `response` (object): Complete service response, including pagination.

**Example output**

```json
{
  "response": {
    "api_version": "example",
    "drive_id": "example",
    "drive_name": "example"
  }
}
```

### Create Import URL

Create Import URL.

**Inputs**

- Partner drive id (`partner_drive_id`, text, required): The drive id associated with the auth token.
- Project schema (`project_schema`, object, required): The Project schema

**Outputs**: Result returned by the service.

- `response` (object): Complete service response, including pagination.

**Example output**

```json
{
  "response": {
    "url": "example"
  }
}
```

### Export project transcript

Export project transcript.

**Inputs**

- Project (`project_id`, text, required): The ID of the project to export from.
- Composition id (`composition_id`, text): Composition to export. If omitted, the first composition in the project is used. Accepts any of the following formats: - A full composition UUID (e.g. `39677a40-1c43-4c36-8449-46cfbc4de2b5`) - A 5-character short ID from a Descript URL (e.g. `39677`) - A full Descript project URL (e.g. `https://web.descript.com/{project_id}/39677`)
- Format (`format`, text, required): Transcript file format. The response body is the raw transcript file in the requested format (binary for `docx`, plain text otherwise). The `srt` format exports a SubRip subtitle file with timed captions. One of: `txt`, `markdown`, `html`, `rtf`, `docx`, `srt`.
- Include speaker labels (`include_speaker_labels`, text): Speaker label mode. - `off`: No speaker labels - `changes`: Show speaker label when the speaker changes - `every_paragraph`: Show speaker label on every paragraph. One of: `off`, `changes`, `every_paragraph` (Every paragraph).
- Include markers (`include_markers`, true/false): Include markers in the transcript.
- Timecodes (`timecodes`, object): Timecode options. When provided, timecodes are included in the output.

**Outputs**: Result returned by the service.

- `response` (object): Complete service response, including pagination.

**Example output**

```json
{
  "response": {
    "content_type": "example",
    "file_base64": "example",
    "text": "example"
  }
}
```

### Get job status

Get job status.

**Inputs**

- Job id (`job_id`, text, required): The job ID

**Outputs**: Result returned by the service.

- `response` (object): Status of an async job. The response structure varies based on the job type. Use the `job_type` field to determine which fields will be present.

**Example output**

```json
{
  "response": {
    "created_at": "2026-10-25T12:00:00Z",
    "drive_id": "example",
    "job_id": "example-job",
    "job_state": "stopped",
    "job_type": "agent",
    "progress": {
      "composition_id": {},
      "label": {},
      "last_update_at": {},
      "percent": {},
      "share_url": {}
    },
    "project_id": "example",
    "project_url": "example",
    "result": {},
    "stopped_at": "2026-10-25T12:00:00Z"
  }
}
```

### Get project details

Get project details.

**Inputs**

- Project (`project_id`, text, required): The project UUID

**Outputs**: Result returned by the service.

- `response` (object): Complete service response, including pagination.

**Example output**

```json
{
  "response": {
    "compositions": [],
    "created_at": "2026-10-25T12:00:00Z",
    "drive_id": "example",
    "folder_path": "example",
    "id": "example-id",
    "media_files": {},
    "name": "Example",
    "publishes": [],
    "updated_at": "2026-10-25T12:00:00Z"
  }
}
```

### Get Published Project Metadata

Get Published Project Metadata.

**Inputs**

- Published project slug (`published_project_slug`, text, required): The unique URL slug identifying the published project

**Outputs**: Result returned by the service.

- `response` (object): Metadata for a successfully published Descript project

**Example output**

```json
{
  "response": {
    "download_url": "example",
    "download_url_expires_at": "2026-10-25T12:00:00Z",
    "metadata": {
      "duration_formatted": "example",
      "duration_seconds": 0,
      "published_at": "2026-10-25T12:00:00Z",
      "published_by": {
        "first_name": {},
        "last_name": {}
      },
      "title": "example"
    },
    "privacy": "public",
    "project_id": "example",
    "publish_type": "audio",
    "subtitles": "example"
  }
}
```

### Import media and sequences

Import media and sequences.

**Inputs**

- Project (`project_id`, text): Existing project ID to import media into. If not provided, a new project will be created. When importing into an existing project, media filenames must not conflict with existing files.
- Project name (`project_name`, text): Name for the new project. Only used when project_id is not provided.
- Team access (`team_access`, text): Access level for drive members. Only applicable when creating a new project (when project_id is not provided). Defaults to `none` if not specified. - edit: Users can edit the project - comment: Users can view and comment but not edit - view: Users can view but not comment or edit - none: No shared access (private to owner). One of: `edit`, `comment`, `view`, `none`.
- Folder name (`folder_name`, text): Folder path to place the new project in (e.g. "Clients/Acme/Videos"). Supports nested paths using "/" as separator. Only applicable when creating a new project (when project_id is not provided). Existing folders along the path are reused; missing segments are created automatically.
- Workspace name (`workspace_name`, text): Existing workspace to create the new project in, matched by name (case-insensitive). Only applicable when creating a new project (when project_id is not provided). Reserved names: `Personal` (your private space) and `General` (the shared drive workspace). Any other value is looked up as a custom workspace name; unknown names return 404. When omitted, `team_access` is passed through unchanged. When set to `Personal`, `team_access` must be `none` or omitted. When set to `General` or a custom workspace name, `team_access` must be `edit`, `comment`, or `view`; omitting it defaults to `view`, and `none` is rejected. For custom workspaces, the caller must be a member of that workspace.
- Add media (`add_media`, object): Map of media reference IDs (display names with optional folder paths) to media import items. Keys are the display names that will appear in the project (e.g., "Misc/intro.mp4" or "demo.mp4"). Values define how to import each media item (URL import or multitrack sequence).
- Add compositions (`add_compositions`, list): Optional list of compositions to create in the project
- Callback url (`callback_url`, text): Optional webhook URL to call when the job completes or fails. Descript will POST the job status (same format as [GET /jobs/{job_id}](#operation/getJob)) to this URL.

**Outputs**: Result returned by the service.

- `response` (object): Response returned when creating an import job

**Example output**

```json
{
  "response": {
    "drive_id": "example",
    "drive_name": "example",
    "job_id": "example-job",
    "project_id": "example",
    "project_url": "example",
    "upload_urls": {}
  }
}
```

### List agent models

List agent models.

**Inputs**: None.

**Outputs**: Result returned by the service.

- `response` (object): Complete service response, including pagination.

**Example output**

```json
{
  "response": {
    "aliases": [],
    "availableModels": []
  }
}
```

### List jobs

List jobs.

**Inputs**

- Project (`project_id`, text): Filter by project ID
- Type (`type`, text): Filter by job type. One of: `import/project_media` (Import/project media), `agent`.
- Cursor (`cursor`, text): Cursor for the next page of results, obtained from `pagination.next_cursor` in a previous response
- Limit (`limit`, number): Number of items per page (1-100). Defaults to 20. From 1 to 100.
- Created after (`created_after`, date and time): Filter jobs created after this timestamp (ISO 8601). Default: 7 days ago. Oldest allowed: 30 days ago.
- Created before (`created_before`, date and time): Filter jobs created before this timestamp (ISO 8601). Default: now.

**Outputs**: Result returned by the service.

- `response` (object): Complete service response, including pagination.

**Example output**

```json
{
  "response": {
    "data": [],
    "pagination": {
      "next_cursor": "example"
    }
  }
}
```

### List projects

List projects.

**Inputs**

- Name (`name`, text): Filter projects whose name contains this string (case-insensitive).
- Folder path (`folder_path`, text): Filter projects by folder path (e.g. "Clients/Acme/Videos"). Use "/" to separate nested folders. Returns only projects directly inside the deepest folder.
- Created by (`created_by`, text): Filter projects created by this user UUID. Pass `me` to filter by the authenticated user.
- Created after (`created_after`, date and time): Filter projects created after this ISO 8601 timestamp.
- Created before (`created_before`, date and time): Filter projects created before this ISO 8601 timestamp.
- Updated after (`updated_after`, date and time): Filter projects updated after this ISO 8601 timestamp.
- Updated before (`updated_before`, date and time): Filter projects updated before this ISO 8601 timestamp.
- Sort (`sort`, text): Sort field. Defaults to created_at. One of: `name`, `created_at` (Created at), `updated_at` (Updated at), `last_viewed_at` (Last viewed at).
- Direction (`direction`, text): Sort direction. Defaults to desc. One of: `asc`, `desc`.
- Cursor (`cursor`, text): Pagination cursor from a previous response's `pagination.next_cursor`.
- Limit (`limit`, number): Number of projects per page (1-100). Defaults to 20. From 1 to 100.

**Outputs**: Result returned by the service.

- `response` (object): Complete service response, including pagination.

**Example output**

```json
{
  "response": {
    "data": [],
    "pagination": {
      "next_cursor": "example"
    }
  }
}
```

### Publish project media

Publish project media.

**Inputs**

- Project (`project_id`, text, required): The ID of the project to publish.
- Composition id (`composition_id`, text): Composition to publish. If omitted, the first composition that has content is used, skipping the empty placeholder that leads projects created by an agent or import job. Accepts any of the following formats: - A full composition UUID (e.g. `39677a40-1c43-4c36-8449-46cfbc4de2b5`) - A 5-character short ID from a Descript URL (e.g. `39677`) - A full Descript project URL (e.g. `https://web.descript.com/{project_id}/39677`)
- Media type (`media_type`, text): Media type of the published output. Defaults to `Video` when omitted. If the target composition has no video content: - omitting `media_type` publishes it as `Audio` (the completed job result reports `media_type: Audio`), - explicitly requesting `Video` is rejected with a 422. One of: `Video`, `Audio`.
- Resolution (`resolution`, text): Resolution for the published output. Only applicable when media_type is Video. One of: `480p`, `720p`, `1080p`, `1440p`, `4K`.
- Callback url (`callback_url`, text): Optional webhook URL to call when the job completes or fails. Descript will POST the job status (same format as [GET /jobs/{job_id}](#operation/getJob)) to this URL.
- Access level (`access_level`, text): Desired access level for the published share page. If omitted, the drive's configured default is used. Returns 403 if the requested level is not permitted by the drive's publish settings (e.g. requesting `public` when search engine indexing is disabled). One of: `public`, `unlisted`, `drive`, `private`.

**Outputs**: Result returned by the service.

- `response` (object): Complete service response, including pagination.

**Example output**

```json
{
  "response": {
    "drive_id": "example",
    "drive_name": "example",
    "job_id": "example-job",
    "project_id": "example",
    "project_url": "example"
  }
}
```

### Search a drive

Search a drive.

**Inputs**

- Query (`query`, text, required): Search term. Matched against names, contents, and visual matches (on drives with the Enterprise plan). Must be non-empty.
- Updated after (`updated_after`, date and time): Return results updated at or after this time. Accepts an ISO 8601 date (`2026-08-01`, interpreted as the start of that UTC day) or timestamp (`2026-08-01T09:30:00Z`). Values without a timezone offset are read as UTC.
- Updated before (`updated_before`, date and time): Return results updated at or before this time. Accepts an ISO 8601 date (`2026-08-01`, interpreted as the end of that UTC day) or timestamp (`2026-08-01T23:59:59Z`). Values without a timezone offset are read as UTC.
- Owner (`owner`, list): Return items owned by these user UUIDs. Repeat this parameter to include more than one owner. If omitted, results from all owners are returned.
- Type (`type`, list): Result types to search. Repeat this parameter to search more than one type. If omitted, all result types are returned. - `project`: Projects - `layout_pack`: Layout packs - `project_folder`: Folders that hold projects - `media_library_folder`: Folders in the drive media library - `video`, `image`, `audio`: Files of that media type in the drive media library, Brand Studio, and inside projects. One or more of: `project`, `video`, `image`, `audio`, `project_folder` (Project folder), `media_library_folder` (Media library folder), `layout_pack` (Layout pack).
- Match (`match`, list): How the query may match. Repeat this parameter to allow more than one kind. `name` matches project, file, folder, and layout pack names. `content` matches transcripts and composition text. `visual` matches what appears in videos and images, and is only available on drives with the Enterprise plan. `visual` cannot be combined with another match kind. If omitted, only names and contents contribute. One or more of: `name`, `content`, `visual`.
- Sort (`sort`, text): How to sort results: - `relevance`: closest matches first. This is the default. - `newest`: most recently modified first. - `oldest`: least recently modified first. One of: `relevance`, `newest`, `oldest`.
- Limit (`limit`, number): Maximum number of results to return. Defaults to 30. Maximum is 100. From 1 to 100.

**Outputs**: Result returned by the service.

- `response` (object): Ranked search results.

**Example output**

```json
{
  "response": {
    "results": []
  }
}
```
