Revolutionize transcription with unmatched accuracy, speed, and language support.
27M visits/mo+0.9%#4 Audio & Voice
About Resum
Resum provides an API and SDK for capturing live conversations and processing them with named agents. It opens a speech socket that captures audio from a microphone, file, or live stream. The system saves AAC audio bytes and appends finalized turns to a transcript while the session is active. Once the session ends, each named agent generates a file containing notes, summaries, or other structured outputs. The session follows a fixed workflow: opening a socket, capturing audio, processing turns, and writing agent files after termination. The terminal runner and API share this workflow, with control handled via commands or time limits. Audio is streamed in ten-second chunks at 16 kHz mono, and agent files are written using GPT-4o-mini after the session concludes. The tool requires specific API keys for operation and agent processing but does not log secret values.
Key features
- Live audio capture from microphone, file, or RTMP stream
- Real-time transcript generation with finalized turns
- Named agent processing (notes, summary, etc.)
- AAC audio capture saved as it is streamed
- Session control via commands or time limits
- Environment-based API key authentication
- Structured output files per agent
- Terminal runner and API share the same session workflow
Use cases
- Live meeting transcription and note-taking
- Customer support call analysis and summarization
- Podcast or interview processing with agent-generated notes
Pros
- Captures live audio from multiple sources (microphone, file, RTMP stream)
- Processes conversations in real time with finalized turns appended to transcript
- Generates structured agent files (notes, summaries) after session ends
- Supports multiple named agents per session
- Does not store secret keys in logs
Cons
- Available only to enterprises, not self-serve
- Requires sales contact for pricing and licensing
- Limited to real-time processing with AssemblyAI integration
Frequently asked questions about Resum
What does Resum do?
Resum provides an API and SDK for capturing live conversations and processing them with named agents. It opens a speech socket to capture audio from a microphone, file, or live stream, saves the audio, and generates structured outputs like notes or summaries once the session ends.
Who is Resum designed for?
Resum is designed for enterprises that need to process live conversations in real time. It is not available for self-serve purchase and requires an agreement with Viroke Technologies Inc.
How does Resum process audio during a session?
Resum streams audio in ten-second chunks at 16 kHz mono. The audio is saved as AAC bytes and sent to AssemblyAI for transcription, while finalized turns are appended to a transcript in real time.
What happens after a session ends?
Once the session ends, each named agent generates a file containing structured outputs such as notes or summaries. These files are written using GPT-4o-mini and include details like excerpts, action items, or full summaries.
What API keys are required to use Resum?
Resum requires a STT_KEY to open and run a live session. An OPENAI_API_KEY is needed only when agent files are written after the session concludes. Secret values are not logged.
How do I get started with Resum?
To get started, contact Viroke Technologies Inc for an enterprise agreement. The SDK can then be used to open a session, capture audio, and generate agent files after the session ends.
