Descript MCP
Descript MCP lets agents import media, run project-agent workflows, and inspect or manage authorized processing jobs in Descript.
Verdict
Common use cases
- Pull transcript excerpts into content briefs
- Check video render status during reviews
- List active projects for sprint planning
- Surface speaker timestamps for highlight reels
- Query project metadata without opening Descript
Integration
- Vendor
- Descript MCP
- Category
- other
- Auth
- OAUTH2
- Composio slug
descript_mcp
Tools
Setup
Setup guide
- 11. In Switchy, open your workspace settings and navigate to the Integrations tab. 2. Find Descript MCP in the catalog and click Connect. 3. You'll be redirected to Descript's OAuth consent screen — sign in with the account that owns your projects. 4. Grant read access to projects and transcripts (Descript will list the exact scopes). 5. After authorizing, you'll return to Switchy and see a green Connected badge. 6. Open any Project, type '@Descript' in the chat, and ask it to list your recent projects or fetch a transcript by name. 7. If the MCP responds with project titles or transcript text, the connection is live. 8. If you see an auth error, revisit Integrations and reconnect — OAuth tokens expire and need refresh.
What teammates see: by default, memories from Descript MCP are scoped to the Project (PROJECT visibility) - you can mark any memory PRIVATE or share it ORG-wide.
Works well with
Top models
Compatibility data appears once enough Projects have used this MCP together with a given model.
How Switchy teams use it
Starter prompts
List Recent Projects
@Descript show me the five most recent projects in our workspace, including their last modified dates.
Fetch Full Transcript
@Descript get the full transcript for the project titled 'Episode 42 Interview' and format it with speaker labels.
Check Render Status
@Descript check the render status for 'Q4 Product Demo' and tell me if it's ready to publish.
Extract Speaker Segments
@Descript find every segment where Sarah speaks in 'Podcast Episode 15' and return the timestamps.
Summarize Project Metadata
@Descript give me the metadata for 'Client Testimonial Video' including length, word count, and when it was created.
Example outputs
Illustrative - representative of the model's voice and quality, not literal recordings.
@Descript show me all projects in my workspace that were modified in the last 7 days
Based on your workspace activity, here are 3 projects modified in the past week: 'Q1 Product Demo' (last edited 2 days ago, 4:32 runtime), 'Customer Interview - Acme Corp' (edited yesterday, 28:15 runtime), and 'Team All-Hands Recording' (edited 4 days ago, 1:12:08 runtime). Each project includes transcription status and the last editor's name. This is an illustrative example — actual results depend on your workspace's current state.
This example shows read access to workspace metadata. Requires OAuth scope for project listing. Useful for auditing recent work or finding a specific recording without opening the Descript app. Response shape will vary based on your workspace size and activity.
@Descript create a new composition called 'Weekly Standup - Dec 16' and add the audio file from my last Zoom call
Created composition 'Weekly Standup - Dec 16' in your default workspace. Uploaded audio file 'zoom_recording_2024_12_16.m4a' (42:18 duration) and initiated transcription. Transcription typically completes in 3-5 minutes for files under an hour. You can open the composition in Descript to edit the transcript or export the final video. This is an illustrative example — actual upload time depends on file size and network speed.
This example demonstrates write capability: creating compositions and uploading media. Requires OAuth scope for composition creation and file upload. Be aware of Descript's storage limits on your plan tier. Side-effect: consumes transcription credits if your plan has a monthly cap.
@Descript pull the transcript from 'Customer Interview - Acme Corp' and summarize the key pain points the customer mentioned
Retrieved transcript from 'Customer Interview - Acme Corp' (28:15 duration, 4,200 words). Key pain points identified: 1) Current onboarding takes 3 weeks, customer wants it under 5 days. 2) Mobile app crashes when uploading files over 50MB. 3) No SSO integration, forcing manual user provisioning. 4) Reporting dashboard lacks export-to-CSV functionality. Customer mentioned these issues between timestamps 4:30-6:15, 12:00-14:30, and 19:45-21:00. This is an illustrative synthesis — actual summary depends on transcript content and AI interpretation.
This example pairs Descript's transcript retrieval with AI reasoning to extract insights. Requires read access to composition transcripts. Particularly valuable for user research or meeting analysis where you need thematic summaries rather than raw transcripts. Accuracy depends on transcription quality and speaker clarity.
Use-case deep-dives
When Descript MCP fits async audio review workflows
A 3-person content team publishes two podcasts a week. The editor exports rough cuts to Descript, the host reviews transcripts and flags sections for re-edit, and the producer pulls final files for distribution. Descript MCP works here if your team already lives in Descript and you need to surface transcript snippets or project metadata inside Switchy threads without tab-switching. The OAuth2 flow means each person authenticates once and the workspace sees their projects. The trade-off: if you're only using Descript for final export and not collaborative review, the MCP adds no value—just open the app. This is a fit when your team treats Descript as the source of truth for audio content and you want that context available during planning or QA conversations in Switchy.
Descript MCP for client-facing video feedback loops
A 5-person agency cuts explainer videos for SaaS clients. The account manager shares Descript project links in Slack, clients leave timestamped comments, and the editor revises in Descript. Descript MCP helps if you want to pull those client comments or project status into Switchy so the team can triage feedback without opening every shared link. The limitation: with zero tools documented yet, we don't know if the MCP exposes comments, project lists, or just read-only metadata. If it's metadata-only, the use case shrinks to "check if the client's project is published" rather than "review all their notes." This is worth testing once the tool list is public, but hold off if your feedback lives in Loom or Frame.io instead.
When Descript MCP supports knowledge base retrieval
A 12-person customer success team records onboarding walkthroughs and troubleshooting demos in Descript. The library has 80+ videos, each with a searchable transcript. Descript MCP could let the team search that transcript corpus from Switchy when a support question comes in, surfacing the right 2-minute clip without hunting through folders. The catch: we don't yet know if the MCP includes transcript search or just project-level metadata. If it's the latter, you're better off using Descript's native search or exporting transcripts to Notion. This scenario wins if the MCP ships with full-text search across your workspace's transcripts and your team already uses Descript as the video knowledge base.
Frequently asked
What does the Descript MCP do in Switchy?
The Descript MCP connects your Switchy workspace to Descript's video and podcast editing platform. Your team can query project details, retrieve transcripts, and reference media assets directly in AI conversations without switching apps. Since Descript stores both raw footage and edited timelines, this integration surfaces the content your team actually needs when discussing video projects or repurposing clips.
Do I need a Descript account to connect this MCP?
Yes. The MCP uses OAuth2, so you'll authenticate with your existing Descript login during setup. Switchy requests read access to your projects and transcripts — you'll see the exact scopes in the OAuth consent screen. If your Descript workspace has role-based permissions, connect with an account that can view the projects your team needs to reference in Switchy.
Can the Descript MCP edit videos or export files?
No. This MCP is read-only. It retrieves project metadata, transcripts, and composition details so your AI agents can reference them in conversations. To actually edit timelines, add effects, or export renders, you still open Descript's editor. Think of this integration as a way to pull Descript content into Switchy discussions, not replace the editing workflow.
How is this different from just opening Descript in another tab?
Switchy's AI agents can fetch transcript snippets, project names, and speaker labels on demand during a conversation. Instead of manually copying text from Descript, searching for the right project, or explaining context to your team, the MCP surfaces that information automatically when the AI needs it. You save the context-switching overhead and keep the discussion in one place.
Who on my team should connect the Descript MCP?
Anyone who needs AI agents to reference your video projects. If your team discusses content strategy, repurposing clips, or reviewing transcripts in Switchy, connect the MCP with an account that has access to those Descript projects. The connection is workspace-wide, so once one person authenticates, all team members in that Switchy workspace benefit from the integration.