Remote MCP · OAuth
Integrations
Move YouTube transcripts into the tools you already use.
Connect Threadmap directly to an AI agent through remote MCP, or call the transcript API from visual automation platforms and your own software.
HTTP Request node
n8n
Trigger transcript extraction from an n8n workflow and pass structured text into databases, AI nodes, or publishing systems with source timestamps intact.View n8n setup →HTTP module
Make
Add YouTube transcript extraction to a visual Make scenario with one authenticated HTTP request and reusable structured output for later modules.View Make setup →Webhooks action
Zapier
Send a YouTube URL from any Zap trigger to Threadmap and continue the automation with structured, timestamped transcript data and clear request metadata.View Zapier setup →Two ways to connect
Agent-native MCP or a predictable REST API.
Codex and compatible AI clients can discover Threadmap tools, authenticate through OAuth, and retrieve the source transcript when a conversation needs it.
Read the MCP connection guide →n8n, Make, Zapier, backend services, and scripts can send a video URL and receive structured JSON with text and timestamped segments.
Read the API documentation →Build a reliable pipeline
Extract once, then use the transcript anywhere.
Threadmap handles transcript retrieval and optional audio fallback. Your workflow decides what happens next: save the source, index it for search, summarize it with an AI model, generate accessible notes, compare videos, or notify a team. Keeping extraction separate from downstream processing makes every step easier to inspect and replace.
Structured output
Receive full transcript text plus timestamped segments, language, duration, source mode, and request metadata.
Server-side keys
Bearer authentication works in automation tools and application servers. Never expose a live key in public browser code.
Clear usage
API response headers report the remaining non-expiring minute balance for monitoring and alerts.
One private library
Transcripts created by an account can be saved into the same searchable library used by playlist imports.
Example automations
Start with a concrete job, then add complexity.
Each integration uses the same transcript engine and account balance. Begin with one URL and inspect the returned fields before processing a playlist or connecting another service.
Receive a saved YouTube link, extract its transcript, store the source URL and text, and notify a researcher that searchable material is ready.
Turn selected interviews into timestamped source documents before a separate AI step proposes themes, quotations, outlines, or derivative content.
Process public lesson videos and deliver readable text beside the original material. Preserve timestamps so learners can move between reading and watching.
Index transcript segments with video IDs and start times. Search results can point users back to the precise moment where information was spoken.
Route errors and low-confidence source cases to a person instead of publishing automatically. The request ID helps support trace a failed call without exposing credentials.
Start connecting
Choose your workflow and use one Threadmap balance.
Minute packs work across API, MCP, and playlist imports. Failed requests do not spend successful transcript minutes, and usage can be checked before large jobs.