# Resum > Resum connects live audio to named AI agents, producing a real-time transcript and agent-specific files when the session ends. It is for developers who want to integrate conversational AI into their own applications via a simple SDK. - Type: AI tool - Page: https://completeaitraining.com/ai-tools/resum/ - Published: 2026-11-03 - Updated: 2026-11-05 - Author: Jeroen Erne - Jobs: IT and development - Topics: Developer Tools, API, Speech-To-Text - Reading time: 2 min ### About Resum Resum is an API and SDK for live conversations, launched as an enterprise voice agent infrastructure tool. It connects a live conversation to named AI agents you define, producing a transcript during the session and one file per agent when the session ends. The SDK wraps the session lifecycle into five calls: create, wait until ready, pipe audio in, listen while live, and end. ### Review Resum approaches live voice processing with a narrow, agent-oriented structure. Instead of a general-purpose transcription pipeline, it routes a single audio stream to multiple named agents and gives each one its own output file. The tool is brand new-launched this week-so documentation and community patterns are still forming. #### Key Features - Named agent routing-each agent you name (e.g., "notes", "summary", or any custom name) receives its own output file when the session ends - Fixed-shape outputs for built-in agent names: "notes" and "summary" produce structured results, while any other name you pass returns plain text - Live transcript building-finalized turns append to a transcript while the session is still open - Session-end flush-stopping the session finishes all pending work, flushes remaining audio, and writes agent files before closing - Five-call SDK sequence: create a session, wait for the speech socket to be ready, pipe audio from a microphone/file/live stream, listen during the session, and end it #### Pricing and Value Pricing is not yet defined on the product page. The tool is described as an enterprise API and SDK, and the maker invites direct contact at williams@viroke.com for those who want to run it with their own conversations. No free tier, trial, or usage-based pricing details have been published. #### Pros - Single audio stream feeds multiple agents simultaneously, reducing the need for parallel processing setups - Audio stays local until the speech socket opens, which gives you control over when transmission begins - The same recording is saved exactly as sent-no pre-processing alters the input before agent distribution - End-session flush guarantees agent files are complete before the session terminates, a detail noted by early commenters as useful for downstream automation #### Cons - No pricing information is available, making cost evaluation impossible without contacting the maker directly - Documentation is minimal-the launch announcement covers the core flow, but error handling, retry logic, and edge cases aren't detailed yet - The tool is not well suited for teams that need a managed, browser-based transcription service with a GUI; it's an API/SDK aimed at developers comfortable integrating at the code level Resum fits scenarios where a single conversation needs to feed multiple downstream processes with different output formats-interview platforms generating separate candidate notes and interviewer summaries, for example. Developers who want programmatic control over when audio transmission starts and stops will find the socket-ready gating useful. Teams looking for a turnkey, UI-driven transcription product should look elsewhere, as this is a developer API with no graphical interface. --- This is the Markdown form of https://completeaitraining.com/ai-tools/resum/. The whole site is mapped at https://completeaitraining.com/llms.txt.