High-quality audio and video
Share pre-recorded audio and video, or a screen or application window with remote participants without the need to download or install an extension. All SDKs include embeddable screen viewing controls that provide the ability to zoom in and out or pan on the shared content. Callbacks track who is sharing and when the source of the sharing changes.
Track the active speakers through a periodic callback with a list of attendees sorted from most to least active. Use this information to enrich the application user experience with active talker labels, video tile highlighting or prioritization, and top talker information.
Volume notifications and mute controls
Use real-time notification callbacks to provide volume levels of attendees who are actively contributing audio for the session. Mute and unmute microphone actions and notifications are available to indicate the mute status for each attendee.
Use the Device Controller module to manage the available audio input, audio output, and video input. Control permissions, preview, and selection of the devices used during the media session.
Manage audio and video performance with both pre-built and customizable bandwidth policies controlling video bandwidth adaptation and video stream pausing. Builders can override these policies to get a more customized experience. The audio stream auto-reconnect feature provides more resilience to network interruptions.
Deliver real-time content over WebRTC to 10,000 audience members. Ideal for large meetings, webinars and lectures where audience members may want to interact with the presenters. The primary WebRTC media session can be replicated to up to 40 additional WebRTC media sessions for read-only consumption by audience members. Audience members can seamlessly transition to and from the primary session, without any WebRTC reconnection, to interact with presenters.
WebRTC sessions include integrated Traverse Using Relay around NAT (TURN) service for corporate firewall and NAT transversal.
Workflows with Amazon EventBridge events
Leverage Amazon EventBridge events to trigger workflows based on session events like starting a call queue workflow when the session is created or when a specific attendee joins or leaves.
Public Switched Telephone Network (PSTN) Audio
Phone Numbers for voice applications and meetings
Create or port your personalized dial-in numbers for the Amazon Chime SDK - based telephony applications and meetings. Toll-free and toll numbers are supported. You can also bring your own carrier and phone numbers and connect to your existing carrier using Amazon Chime Voice Connector (Session Initiation Protocol) SIP Trunking.
Build voice bots using Amazon Lex so that callers can use their voice to navigate phone system menus. You can build serverless phone self-service solutions and directly access Amazon Lex bots using a phone number provisioned within Amazon Chime SDK. The Amazon Chime SDK PSTN audio integration with Amazon Lex enables you to add natural language voice conversational experiences in your commercial or custom applications. You can also use Amazon Chime Voice Connector and Amazon Chime SDK PSTN audio to access Amazon Lex bots in any enterprise communications or contact center application that uses Session Initiation Protocol (SIP).
Add call recording and post-call transcription to your voice applications. Using an API, you can record call audio and store the recording files in your own Amazon Simple Storage Solution (Amazon S3) bucket.
SIP Trunking Setup
Amazon Chime Voice Connector supports standards-based Session Initiation Protocol (SIP); it is tested for compatibility with common SIP-based Private Branch Exchanges (PBXs), call centers, and Session Border Controllers (SBCs). With Amazon Chime Voice Connector, you can start using your SIP trunks in minutes using the Amazon Chime console or AWS SDK. There are no service orders needed and no trunk set-up charges.
T.38 Fax Support
Send and receive faxes from traditional fax machines connected to your phone system or via your SIP-enabled fax server.
Implement Call Recording and Analytics
Capture real-time audio from SIP trunking and SIP-based media recording (SIPREC). Build real-time and post-call analytics using Amazon Transcribe, Amazon Comprehend, and other Machine Learning (ML) services.
Proxy Phone Sessions
Enable communications between two users via PSTN, without sharing personal phone numbers. Amazon Chime SDK proxy phone sessions support two-way voice or text communications sessions for up to 12 hours while protecting user privacy and keeping transactions within your service.
Phone Number Management and Porting
Manage your phone number inventory, order new numbers, and review pending transactions using the Amazon Chime console, AWS SDK, and AWS CLI. Keep your existing phone numbers by having us port them from your previous service provider. For more information please refer to Managing phone numbers in Amazon Chime SDK.
Call Health Monitoring
Use Amazon CloudWatch to get a unified view of voice calling operational health that includes call completion rates and voice quality metrics. You can use CloudWatch to detect anomalous behavior, set alarms, visualize logs and metrics, automate actions, and troubleshoot issues to keep your voice calling running smoothly.
Security and Authentication
You can use encryption to secure calls between your phone system and AWS. Amazon Chime Voice Connector supports TLS to encrypt SIP signaling and SRTP to encrypt media streams. You can optionally enable SIP digest authentication to authenticate outbound calls from your phone system.
Quality and Reliability
Amazon Chime Voice Connector supports fault-tolerant, cross-region inbound and outbound call routing to and from your phone system. In the event of loss of connectivity between an AWS region and your phone system, or if there is an issue with the Amazon Chime Voice Connector service in a region, calls will route to the other Amazon Chime Voice Connectors.
Cost-effective Network Access
You can access your Amazon Chime Voice Connector over AWS Direct Connect or your Internet connection. Using AWS Direct Connect, you can establish private connectivity between AWS and your data center, office, or colocation environment, which can help provide a more consistent network experience than Internet-based connections. With the elastic scaling, you can meet all your call volume needs, including unexpected or seasonal spikes, and you no longer need to purchase bursting capacity or over-provision concurrent call paths.
Channel flows enable you to process messages in channel processors (AWS Lambda functions) before they are delivered. Channel flows can be used to remove sensitive data such as government ID numbers, phone numbers, or profanity from messages before they are delivered. Channel flows can also be used to perform functions like aggregation of responses to a poll before sending results back to participants or sending messages to recipients via SMS or other popular messaging applications.
Create public channels that any user can find and join, or private channels that require members or moderators to add new members.
Message history is available to your users for as long as you choose.
Allow moderators and users to redact messages containing sensitive data so they are no longer accessible by users.
Enable users to edit any message they send.
Using the Amazon Chime SDK, user channel moderators can redact inappropriate messages, kick users out of channels, and ban users to prevent them from rejoining.
Include attachments stored in Amazon S3 or your choice of storage location as message metadata.
Implement retention policies, if required. Automatically delete any messages after a retention period you choose, ranging from 1 day to 15 years. APIs are also available to delete individual messages and channels, permanently removing them from Amazon Chime.
There are no hard limits on the number of users that can use Amazon Chime SDK messaging.
Choose to use Amazon Cognito for user management and authentication, or use your existing identity provider.
Use control messages, up to 30 bytes, for features like typing indicators, message effects, or presence.
Export Amazon Chime SDK chat data through Amazon Kinesis to provide search, archive data, process data, or train ML models.
Use React UI components for common messaging elements.
Search for conversations by members or unique ID such as an appointment ID or game ID.
Automatically pre-fetch information to display a rich chat UI to users when clients load including surfacing channels that require the users attention first.
Capture the contents of your Amazon Chime SDK meeting and save them to an Amazon Simple Storage Service (Amazon S3) bucket of your choice. Media capture allows you to capture meeting audio, video, and content share streams in five second segments and directly deliver them to your designated Amazon S3 bucket.
Capture a single stream of the active speaker video along with the combined meeting audio, or to capture individual video tile streams with a separate combined meeting audio stream, or only the combined meeting audio. When capturing individual video streams, use the SourceConfiguration to select specific attendee video streams to capture.
In addition to media, meeting events and data messages are also delivered to your designated Amazon S3 bucket. When live transcription is active in the meeting, transcription data is also delivered to the S3 bucket for post meeting processing.
Noise reduction with Amazon Voice Focus
Reduce unwanted background noise with Amazon Voice Focus, which uses deep learning and models of speech and hearing to help keep environmental noises like wind or running water from interfering with conversations. Amazon Voice Focus is designed to reduce background noises such as dogs barking as well as foreground noises like keyboard typing to help listeners focus on the primary interaction.
Echo reduction with double-talk improvement
Disruptive acoustic echoes can happen in a teleconference when the sound played by the loudspeaker at one end of the call is picked up by the microphone and recirculated back into the conversation. Double-talk occurs when parties at different ends of an audio call are speaking at the same time. Our echo reduction technology uses machine learning to help keep acoustic echoes from recirculating while preserving the quality of the desired speech even in double-talk conditions.
Live transcription with Amazon Transcribe
Use live transcriptions generated by Amazon Transcribe or Amazon Transcribe Medical to overlay subtitles, build a transcript, or perform real-time content analysis. Individual audio streams of the top talkers are sent directly to your Amazon Transcribe account for real-time transcription. The results are processed into user-attributed transcription information which is then sent to every meeting participant via Amazon Chime SDK data messages. Transcription information can also be stored in your Amazon Simple Storage Service (Amazon S3) bucket with media capture for post-meeting processing.
Video background replacement or blur
Help users increase their visual privacy with video background replacement or blur. Each user’s webcam video is processed on their device with machine learning to segment the background from the user. Background replacement inserts an image behind the user, replacing the detected background. Alternatively, the background can be blurred with a high-strength blur to help increase privacy, or a low-strength blur to create a bokeh effect.
Users can confirm their background replacement or blur prior to joining a meeting through video preview. The segmentation model for separating the user used to separate a user from their background can be substituted with other compatible models designed for specific applications or performance goals.
Amazon Voice Focus Amazon Machine Image (AMI)
Amazon Voice Focus AMI is an Amazon Linux 2 AMI (x86_64) with an embedded Amazon Voice Focus library along with an Amazon Voice Focus demo executable and example scripts. Amazon Voice Focus helps reduce noise in real-time speech capture or archived speech recordings. With no up-front costs or commitments required, customers simply pay hourly for each Amazon Voice Focus AMI instance they use.
To get started with Amazon Voice Focus AMI, please fill out this form.