MCP-native
Built as a first-class Model Context Protocol server, so any MCP client can discover and call the tools with no glue code. OAuth handles authorization end to end.
Audiogram API handles the catalog, the audio, and the transcription so you can focus on what your agent does with the words.
Built as a first-class Model Context Protocol server, so any MCP client can discover and call the tools with no glue code. OAuth handles authorization end to end.
The same operations are exposed as a clean REST API with scoped API keys, so your services can search and fetch transcripts without an agent in the loop.
Search the full public catalog by show or episode. Results carry in_catalog and addable flags so agents know exactly what to fetch versus ingest.
Ingest an episode and we download the audio and queue transcription for you. Poll get_transcript and receive clean text when it is ready.
Per-operation limits are surface-agnostic and tied to your account, not the client. Search is generous, ingest is metered monthly, and limits scale with your plan.
Every call is recorded as a usage event. The console shows what you have spent across MCP and REST so there are no surprises at the end of the month.
Find shows by name across the Apple Podcasts catalog, with feed URLs ready for ingestion.
Search episodes and get in_catalog / addable flags so agents know what to fetch versus add.
Ingest an episode by id. Audio is downloaded and queued for transcription automatically.
Pull transcript text by episode id, with a polite poll-and-wait while transcription runs.