SentDm.Voice.Agents 0.2.0

Sent Voice Agents for .NET

An AI agent that takes and places phone calls through Sent, on the Sent Voice Agents protocol 1.0.0, for .NET 8 and later.

Status: Preview, with the protocol. Generated and rendered by sent-dm-sdk-pipeline from the protocol at v1.0.0; every release passes the protocol's conformance scenarios in CI.

Read the Voice Agents docs to get started.

Install

dotnet add package SentDm.Voice.Agents

.NET 8 or later. The package depends on nothing outside the runtime: System.Net.WebSockets carries the connection and System.Text.Json the frames.

Quickstart

using SentDm.Voice.Agents;

class Support : Agent
{
    public Support(AgentOptions options) : base(options) { }

    public override Task OnSessionStartAsync(Session session)
    {
        session.Say("Hi, this is Acme support. How can I help?");
        return Task.CompletedTask;
    }

    public override async Task OnUserSaidAsync(Session session, UserSaid said)
    {
        if (said.Text.Contains("person"))
        {
            session.Say("Transferring you now.");
            await session.TransferAsync("+14155550100");
        }
        else
        {
            session.Say($"You said {said.Text}");
        }
    }
}

await new Support(new AgentOptions
{
    ApiKey = Environment.GetEnvironmentVariable("SENT_API_KEY")!,
    AgentId = "agent_...",
}).ServeAsync();

ServeAsync() connects to Sent with the account API key and serves calls until SIGINT or SIGTERM; RunAsync() does the same without the signal handlers, and Stop() ends it. Each session's handlers run in order on their own chain, so sessions never wait for one another. The runtime answers heartbeats, opens a new connection when one drains or is lost, and resumes the sessions that can move; the handlers only see calls.

On a Session: Say(text, last, interruptible), Clear(), SendAudio(samples), and the tools TransferAsync(to), HangupAsync(), SendDtmfAsync(digits), HoldAsync(held), each completing with Sent's ToolResult. session.From, To, Direction, SessionType, Language, Context, SampleRate carry what session.start said.

Frames.cs holds every frame as a record, with Protocol.Decode() and Protocol.Encode() that check the protocol's rules; it is generated from the protocol and not edited by hand.

Develop

dotnet tool restore
dotnet build && dotnet test            # the library, the tests against the protocol's fixtures
dotnet csharpier check .               # formatting; Frames.cs is excluded

The conformance scenarios run in CI through the protocol's action, against the SentDm.Voice.Agents.Conformance program. Releases are cut by release-please from main and published to nuget.org from GitHub Actions by trusted publishing, with no stored key.