Talk To My Agent
Tell us about your calls

OpenClaw

Give your OpenClaw agent a phone number.

Our voice gateway installs next to OpenClaw and talks to it over localhost. Callers reach the same agent, with its skills, memory and tools, in a natural voice. And it can place calls too.

Get an install tokenSee setup

Same agent. Now on the phone.

OpenClaw agents live in chat. The gateway adds the phone: it answers your number, does the listening and speaking, and asks your agent whenever the caller needs something only it knows or can do.

Nothing to rebuild. The persona, memory and skills you already have are the ones on the call.

  • Connects toOpenClaw's /v1/responses endpoint, over localhost
  • Runs onThe same Linux or macOS host, as a system service
  • CallsInbound on your number, outbound when your agent places them
  • NetworkNo inbound ports. The gateway connects out through an encrypted tunnel
  • CapacityTwo simultaneous calls by default, raised on request

Setup. One command, same host.

  1. 1

    Have OpenClaw running

    The gateway runs on the same machine and finds OpenClaw on its own.

  2. 2

    Get an install token

    Create one in your dashboard. It works once and expires after 15 minutes.

  3. 3

    Run the installer

    It turns on OpenClaw's Responses endpoint, restarts OpenClaw and installs the gateway as a service.

  4. 4

    Choose a voice and call it

    Pick a voice and a local number in the dashboard, then call your agent.

curl -sSL https://api.talktomyagent.io/install.sh \
  | bash -s -- --token <YOUR_TOKEN> --accept-license

Full walkthrough: setup and configuration.

On a call. What it does in chat, by voice.

Your skills and tools

Mid-call, the gateway asks your agent to look something up or take an action, and speaks the answer.

Knowledge base search

Answers from the documents and data you've given your agent.

Caller memory

With the dojo-relationship-hub CRM skill installed, it recognizes returning callers and saves new ones.

Outbound calls

Ask your agent in chat to call someone. It confirms the number and purpose first. Or place calls through the Voice API.

Long jobs, no hold music

Optional and off by default: slow tasks go to your agent in the background, and the result is posted to your chat channel.

99 languages

The agent answers in the language the caller speaks.

Questions. Specific answers.

What stays on my machine, and what leaves it?

Your agent's tools, memory and data stay on your machine, and it keeps using the model you configured for it. The call itself leaves: our carrier connects it, an encrypted tunnel carries it to your gateway, and Google's realtime voice model does the listening and speaking. That model receives the call audio and the context it needs to answer: your agent's instructions and persona, and what its tools return.

Do I need to open ports or change my firewall?

No. The gateway listens on localhost only and connects out through an outbound encrypted tunnel. Carrier keys never sit on your server.

Are calls recorded?

By default, yes. Recordings play in your dashboard and transcripts come through the API and webhooks. Each is a separate switch. Turn recording off and we keep no audio; turn transcript sharing off and the transcript stays on your host.

Which systems does the installer support?

Linux with systemd, and macOS with launchd. On a Mac it runs as the logged-in user, next to OpenClaw's main agent.

Can it run alongside another agent framework?

If the host also runs Hermes Agent, pass --agent-profile openclaw to the installer to choose OpenClaw. See Hermes Agent for that setup.

Put your agent on the phone.

Get an install token from your dashboard and run one command. Want a hand with your setup? Tell us what you run.

Tell us about your calls