Integrations
Use TwinTone in Muse
TwinTone runs AI hosts — lifelike virtual presenters that livestream for brands and creators. You choose a creator, a product or topic, and a platform; TwinTone sends it live and reports back viewer and usage analytics.
Meta's Muse agent reaches services through connectors. Getting into Meta's reviewed connector directory is a process; the faster route is a user-built custom connector — you ask Muse to build a TwinTone integration straight from TwinTone's API, and it keeps your key in Muse's Secure Credential Store. Nothing is submitted for review, so you can do it today.
The prompt to give Muse
Paste this into Muse, then replace the final paragraph with what you want:
Build me a TwinTone integration and save it as a skill.
TwinTone hosts AI presenters that livestream for brands and creators. Its API
is a small OpenAPI 3.1 REST API.
Use this spec as the connector definition:
https://docs.twintone.ai/openapi.yaml
Production base URL: https://live.twintone.ai/api/v1
Auth: a Bearer API key. Ask me to paste it once and keep it in your Secure
Credential Store — never in a file or a message.
Then let me: list the available AI creators, start a live stream on one of my
connected platforms, check a stream's status, and end it.
Muse reads the spec, builds the tool calls, and saves them as a skill you can reuse. You paste the key once; Muse stores it.
Entry points
| What | Value |
|---|---|
| Base URL | https://live.twintone.ai/api/v1 |
| OpenAPI 3.1 spec | https://docs.twintone.ai/openapi.yaml |
| Discovery | GET https://live.twintone.ai/api/v1 — endpoint index (no key needed) |
Hand Muse the spec URL. It declares the base URL, every endpoint, and the
Bearer auth scheme, so Muse can build the connector with no manual mapping.
Authentication
Create a key at live.twintone.ai/dashboard/api (DEVELOPERS → API Keys), then paste it into Muse once.
| Key | Prefix | Use |
|---|---|---|
| Test | tt_test_ | Free and simulated — the full request/response lifecycle with no real broadcast. Build the connector with this. |
| Live | tt_live_ | Sends real streams to real platforms. |
Test and live keys use the same endpoints and response shapes. A test key
only sees and ends streams started with a test key, never your real ones.
Swapping tt_test_ for tt_live_ is the entire go-live step.
API streams use your plan's streaming minutes; there is no separate API charge. See API pricing.
Honest limits
- Platforms: a stream can go to YouTube, TikTok, Twitch, Instagram, or Facebook. Anything else is rejected.
- Plan gate: test keys are free for any account with a confirmed email.
Live keys need an active Growth or Always-On plan; without one,
creating a live key returns
402 plan_required. - Real streams need a live key. Test keys exercise every call for free but never broadcast.
- Rate limit: 100 requests/min per brand on every endpoint except the
public
GET /creators, which is not rate-limited. Over the limit returns429 rate_limited. - Concurrency: up to 3 streams running at once per brand.
Next steps
Was this page helpful?