Descript for AI agents
Descript imports and processes media through project-agent workflows, letting a team automate video and audio editing jobs instead of doing them by hand. Connect it once through Arc0, and your agent, or Claude, ChatGPT and Cursor, can use it through one MCP endpoint, limited to what each user approved.
What agents do in Descript.
Import media into a project
Import a media file into a project so it's ready for editing or transcription.
Check a processing job's status
Check a job's status before waiting on a long-running export to finish.
Confirm before publishing a project
Publish a project once its edits are reviewed and approved for release.
14 Descript actions, graded by risk.
Every Descript action is tagged read, write or destructive, so one policy covers the whole app and new actions inherit the right default.
read
7Look things up. Allowed by default.
- descript.list_jobsList jobs
- descript.get_projectGet project
- descript.list_foldersList folders
- descript.list_projectsList projects
- descript.get_drive_infoGet drive info
- descript.export_timelineExport timeline
- descript.export_transcriptExport transcript
write
6Create and change things. Allow, or ask the user first.
- descript.import_mediaImport media
- descript.wait_for_jobWait for job
- descript.publish_projectPublish project
- descript.import_drive_mediaImport drive media
- descript.prompt_project_agentPrompt project agent
- descript.report_upload_statusReport upload status
destructive
1Delete, cancel or archive. Ask first, or deny outright.
- descript.cancel_jobCancel job
Descript in three steps.
- 01Your users connect DescriptThey sign in to Descript on Arc0 Connect, under your brand, and approve the access you ask for.
- 02You set the rulesReads run, writes like “import media” can wait for the user, and “cancel job” can be denied outright.
- 03Any agent can actYour agent calls Descript through the Arc0 SDK or MCP, and so can Claude, ChatGPT and Cursor. Every call lands on the audit log.
await arc0.policies.set('descript', { read: 'allow', write: 'ask', // import_media destructive: 'deny', // cancel_job }) # Claude Code: the same connection, one URL $ claude mcp add --transport http arc0 \ https://mcp.arc0.ai/u/u_8f2
How Descript connects.
Descript runs its own MCP server behind OAuth. Arc0 registers the client, users approve access on Arc0 Connect, and your agent reaches Descript through the same endpoint, policies and audit log as every other app.
The same Descript connection serves your agent over MCP and your own backend over REST and the proxy, so a user connects once. How Arc0 handles credentials →
- AUTH
- OAuth 2.0 (MCP)
- CREDENTIALS
- Per-tenant encrypted vault
- MODEL SEES
- Results only, never credentials
- AUDIT LOG
- Every call, on every plan
Use Descript from any agent.
Descript and Arc0, answered.
Can I use Descript with Claude, ChatGPT or Cursor?
Yes. Connect Descript to Arc0 once, then add your Arc0 MCP URL to Claude, ChatGPT, Cursor, Claude Code or any other remote-MCP client. Each assistant only gets the Descript actions you allow.
How do users connect Descript?
Descript runs its own MCP server behind OAuth. Arc0 registers the client, users approve access on Arc0 Connect, and your agent reaches Descript through the same endpoint, policies and audit log as every other app.
Which Descript actions can my agent take?
14 in total: 7 read, 6 write and 1 destructive, such as “import media”. Your policies decide which of them each agent may call.
Can I stop my agent from deleting things in Descript?
Yes. Actions like “cancel job” are graded destructive. Set destructive actions to deny, or to ask so the user approves each one, and blocked calls still show up on the audit log.
Can my own backend call Descript too?
Yes. The same Descript connection is available over REST and through the Arc0 proxy, so your product and your agent share one connection per user.
Plug Descript into your agent.
Your users connect Descript once, under your brand. Your agent gets 14 actions behind your policies, with every call on the record.
Free to build · MCP + REST · Audit log on every plan