Yes: an agent connected to Audioscrape over MCP can call transcribe_audio on a recording, then search_audio and get_transcript on it any time afterward.
Upload the meeting recording to a workspace, name each council member once, and later meetings are transcribed with those names attached automatically.
Search the name on Audioscrape and set an alert on the results; you get an email digest or webhook whenever a newly transcribed episode mentions it.
Open the person's page on Audioscrape for every attributed appearance, then add a speaker filter plus a topic to pull their exact quotes with timestamps.
Use a transcript index with speaker attribution: add a speaker filter plus your terms on Audioscrape, or filter by speaker in the API or MCP tools.
Use a system with persistent speaker identity: name a voice once in your workspace and every later recording is labeled with that name automatically.
Yes: Audioscrape offers a REST API and an MCP server for searching and reading speaker-attributed transcripts of 18M+ podcast episodes with timestamps.
Search a speaker-attributed transcript index; every result carries the speaker, start and end time, and a link that opens the audio at that second.
Add the Audioscrape MCP server at https://mcp.audioscrape.com as a connector; the assistant can then search and read 18M+ podcast transcripts with citations.
29 public recordings, three domains, one scorer for every system. The full grid, including the columns competitors win, and the claims we killed or narrowed before publishing.
We took the same MCP server through OpenAI's app review and Google's Gemini Enterprise connector intake in the same month. One is a human review, the other an automated eval pipeline. Field notes on what each actually checks.
A US–Iran agreement was signed electronically and carried to the G7 as a breakthrough. But the ink was barely dry before the questions — and the skirmishes — started. Here's the story podcasters told.