Bot Overview
Learn about what a bot is, details around customizing bots, and how to access data from the meeting
What is a bot?
A bot is the fundamental entity for accessing a meeting. Bots are single-use and bots are mapped to a single meeting (i.e. you can't have one bot join multiple meeting). The bot itself exists as a participant in the call and has access to all meeting features and data any other participant has.
With a bot, you can interact within the meeting via video, audio, chat, screenshare, and more. You can also access the meeting data (e.g. video, audio, transcription, participants, and metadata) in real-time during a call, as well as after the call for as long as the data is retained. The bot is white label-able, allowing you to customize your bot's appearance and control its interaction in the meeting itself.
Some popular use cases for meeting bots are:
- Streaming/displaying the meeting recording and transcript from the call in your app (real-time or post-meeting) [real time transcription sample app].
- Extracting summaries and intelligence using AI to show in your app (real-time or post-meeting).
- Running analysis/workflows based on the meeting transcript (real-time or post-meeting).
- Having voice/avatar AI agents in meetings who interact with the meeting participants (real-time) [voice agent sample app]
Whether you want to run real-time analysis, transcribe and record meetings, or extract summaries and intelligence using AI, this can all be done through a bot.
Supported platforms
Recall bots support most of the top meeting platforms, but each of the platforms has their own nuances to be aware of. Certain platforms (Zoom, Webex and Slack Huddles) require some setup to start creating bots.
| Platform | Supported | Setup Required? | Setup Guide | Dashboard Link |
|---|---|---|---|---|
| Zoom | ✅ | No | - | - |
| Google Meet | ✅ | No | - | - |
| Microsoft Teams | ✅ | No | - | - |
| Webex | ✅ | Yes | Webex Bot Setup | Webex Credentials |
| Slack Huddle Bots (Beta) | ✅ | Yes | Slack Bot Setup | - |
| Go-To Meeting (Beta) | ✅ | No | - | - |
Creating and scheduling bots
See Creating and scheduling bots for more info.
Customizing your bot
By default, bots show up to meetings like an anonymous participant with their mic/camera off and with the name "Meeting Notetaker".
Regardless of whether you're using our Calendar Integration or the Create Bot endpoint, bots are fully whitelabelable which means you can customize how the bot shows up and how it interacts in the meeting. With bots, you can:
- Customize the bot's name - you can set the bot's name by setting the
bot_namefield in the Create Bot request. - Customize the bot's display image - you can make the bot share an image either through the camera (participant tile) or through a screenshare (if permitted by host settings).
- Make the bot play audio/speak into the meeting using Output Media.
For example, you can make a bot an active participant by turning it into a voice agent and/or live avatar - stream a live avatar into the meeting through the bot's camera and play real-time audio so it can speak/respond.
Accessing meeting data
Bot's have access to all meeting data that regular participants do, saved as recording media. There are two ways to access this data:
- In real-time using Real-Time Endpoints
- After the meeting has completed on the bot recording's media shortcuts
Video + audio
| Mixed | Separate | |
|---|---|---|
| Async (File) | How to get Mixed Video MP4 (Async) | - |
| Real-time (Streams) | How to get Mixed Video + Audio (Realtime) | - |
Video only
| Mixed | Separate | |
|---|---|---|
| Async (File) | (Only accessible as video + audio) | How to get Separate Videos per Participant (Async) |
| Real-time (Streams) | (Only accessible as video + audio) | How to get Separate Videos per Participant (Realtime) |
Audio only
| Mixed | Separate | |
|---|---|---|
| Async (File) | How to get Mixed Audio MP3 (Async) | How to get Separate Audio per Participant (Async) |
| Real-time (Streams) | How to get Mixed Audio (real-time) | How to get Separate Audio per Participant (Realtime) |
Transcription
There are three ways to generate transcripts through Recall:
- Recall.ai Transcription: Using transcription directly through Recall.
- Meeting Caption Transcription: Using the meeting platform's native closed captioning feature to generate a transcript in real-time.
- AI Transcription: Using a 3rd party AI transcription integration.
Which of these you choose depends on your requirements, and we recommend reviewing the documentation for each to decide which best fits your use case.
Participants and participant events
Bots capture several different types of participant data:
- A list of participants who are in the call
- A speaker timeline of all the timestamps of when a different participant started speaking
- A list of all the participant events that occurred throughout the call (e.g. camera on/off, screenshare on/off, mic on/off, chat messages, etc.)
This data is particularly useful for showing the user additional information about the meeting or controlling features based on participant states in your app. See Meeting Participants & Events for more info.
Meeting data
Bots can also capture some data about the meeting itself (e.g. meeting title). See Meeting Metadata for more info.
Updated 1 day ago