elevenlabs say CLI: text to speech from a terminal, and Sume's route

Sume's CLI cannot say a sentence aloud: its docs state that image, video and music generators are not CLI subcommands yet, and the command reference lists no text-to-speech command. A changelog roundup reports that ElevenLabs added a CLI elevenlabs say command on 2026-09-07 with Eleven v3 as its default. To get speech from a terminal on Sume today, call the TTS route over HTTP and read the job result.

This page shows what the CLI does cover and where speech fits.

What does the Sume CLI cover?

The command reference lists account and auth commands, read commands such as sume jobs get, sume jobs result and sume assets list, and sume tools list and sume tools schema for discovering schemas at runtime. So the CLI is useful after a TTS job exists: you can check its status and read its result from the terminal.

How do I speak a line from the shell?

Submit the request with curl, using an Idempotency-Key, then follow the job with sume jobs status and fetch it with sume jobs result. The result lists an audio artifact with a hosted URL that you can download. The default output is MP3 at 44.1 kHz and 128 kbps; ask for wav if you plan to feed the file to an audio tool.

Is there a shorter path from an agent?

Sume's hosted MCP server has a tts_create tool, and the CLI docs point to sume tools list to see what a key can call. For a script or a build step, plain HTTP is the least code. See Jobs and results for the wait modes.

What should I do?

If you only want a quick listen, use the vendor's own CLI or a web playground. If speech is part of a build, wrap the HTTP call in a small shell function and keep the job id with the file. Check the CLI docs again later: the surface changes as new commands ship.