Descript

Descript projects, compositions, media, jobs, drive search and transcripts, plus URL media imports and Underlord agent edits through the Descript API (beta).

Send recordings from storage, CRM or meeting tools into Descript for transcription and editing, track the jobs, and pull transcripts back into content and knowledge flows.

Build a flow with Descript Provider API documentation

Source actions

Destination actions

Data types

Validation status

Implemented · fixture tested.

Implementation is checked against provider-shaped fixtures for every action, authentication failures, rate limits, pagination and duplicate safety. Provider sandbox and authorised account validation are recorded separately when credential-backed runs are completed.

What you can map

agent edits

Run Underlord agent edit
Destination
Not retried automatically
Approval before production
Runs Descript's Underlord agent on a project (or creates one) from a natural-language instruction, for example adding Studio Sound or removing filler words. Returns a job to poll.
agent_edits.create

agent models

List agent models
Source
Models and aliases accepted by agent_edits.create, with relative cost tiers.
agent_models.list

compositions

List project compositions
Source
Compositions (timelines) in a project, with duration and media type.
compositions.list

drive items

Search the Drive
Source
Ranked search across project names, folder names, media file names, composition text and transcripts.
drive_items.search

jobs

Get job status
Source
Status of an import, agent or publish job, with per-file import results and credits used.
jobs.get
List jobs
Source
Incremental
Import and agent jobs from the last 7 days by default (up to 30 days back).
jobs.list

media files

List project media files
Source
Media files in a project (keyed by display path) with type and duration.
media_files.list

media imports

Import media from URLs
Destination
Not retried automatically
Approval before production
Creates a project (or adds to one) from media URLs, optionally builds a composition, and starts transcription. Returns a job to poll.
media_imports.create

projects

Get project
Source
One project with counts of its media files, compositions and existing share pages.
projects.get
List projects
Source
Incremental
Projects in the token's Drive, with name, date and folder filters.
projects.list

published projects

Get published project
Source
Metadata for a share page by its URL slug: title, duration, type, privacy and subtitles.
published_projects.get

publishes

List project share pages
Source
Share pages already published from a project, with URL, access level and dates. Read only.
publishes.list

status

Check API status
Source
Validates the token and returns the Drive it is tied to and the API version.
status.get

transcripts

Export transcript
Source
Exports a composition's transcript as plain text, Markdown, HTML, RTF or SRT subtitles.
transcripts.get

Authentication

Descript personal API token

Plans and access

Any Descript plan with access to API tokens. Imports use the plan's media minutes and agent edits use AI credits. Each personal API token is tied to one Drive and inherits its creator's permissions there. The API is in beta.

Test environment

Descript has no sandbox. Use a separate Drive (and its own token) for testing; imports and agent edits consume real media minutes and AI credits.

Rate limits

Descript returns HTTP 429 with Retry-After, X-RateLimit-Remaining and X-RateLimit-Consumed when a limit is reached; Nexra waits for Retry-After on reads. The published-project endpoint documents 1,000 requests per hour per user. Other per-endpoint limits are not published.

Pagination

Projects and jobs use cursor pagination (pagination.next_cursor, limit up to 100). Search returns at most 100 ranked results in one page.

Provider events

Import, agent and publish jobs accept a per-request callback_url. Nexra does not set it; poll jobs.get instead. Nexra event triggers are not yet available; flows run on demand or on a schedule.

Before you build a flow

The Descript API is in beta. Endpoints, fields and limits follow OpenAPI version 1.2 and may change.
Not implemented on purpose: publishing (POST /jobs/publish creates a share page, which can be public), job cancellation (DELETE /jobs/{job_id}) and any delete. Existing share pages can still be read through projects.get and published_projects.get.
Edit in Descript (POST /edit_in_descript/schema) is a partner API that requires Descript's approval and is not offered.
Direct file uploads are not offered: they need a second binary PUT to a signed URL, which Nexra recipes do not perform. Import media from a URL Descript can fetch (public or pre-signed, supporting HTTP Range requests). YouTube URLs are not accepted by Descript.
Rendered video or audio cannot be downloaded through the API without publishing. Transcripts are available as text, Markdown, HTML, RTF or SRT (DOCX is binary and not offered).
Job history covers the last 7 days by default and at most 30 days.
Each token is limited to one Drive; projects on another Drive return an error.
There is no create, update or rename endpoint for projects or folders other than creating a project through an import or agent edit. No comments, users or analytics endpoints are documented.
media_imports.create and agent_edits.create spend media minutes and AI credits. Both are marked billable, never retried automatically, and Descript answers HTTP 402 when the plan runs out.

Developer reference

Create the Descript credentials described above with the narrowest permissions your flows need.
In Nexra, open Connect, then Connections, then New connection, choose Descript, enter its settings and credentials, and tick only the write actions you need. Connection guide
Test the connection, then use its actions as flow sources and destinations. Flow guide

Descript docsAPI reference Descript docsAuthentication Descript docsRate limits

Machine-readable detail: GET /api/public/connect/connectors/descript and the Connections API.