← Tất cả hướng dẫn
Hướng dẫn Playbook · Miễn phí

Use Groq Models and Whisper Transcription Through Your Muse

Đã xác minh 2026-11-17

Through an AnythingMCP custom connector, your muse can run questions and writing tasks on Groq models with your own key, and turn linked audio files into text with Whisper.

What it does

Groq runs open models fast, including Llama, GPT-OSS, and Qwen for text, and Whisper for speech. Connected this way, the muse gets 9 tools, and none deletes anything. It can list the models your key can use, with owner, context window, and reply limit. It can send a conversation to a model and return the full reply at once. For audio, it can transcribe a file from a public link into text in its own language, with segment or word timing if you ask, and translate speech in a file into English text. It can also list batch jobs started elsewhere and files stored in your Groq organization, without reading their content.

How to use it

  1. Start an AnythingMCP workspace and install the Groq connector.
  2. Sign in to the Groq console, open API Keys, and create a key. Copy it into the connector. It starts with gsk_ and is shown only once.
  3. In Muse, open Settings, then Connectors, and add a custom connector with the address AnythingMCP gives you. Sign in and approve.
  4. Ask the muse to list your models first, so you know which ones your key can use.
  5. Ask in plain words, for example: summarize this email in three points with the small fast model, or transcribe the episode at this link with timings.

What you need

  • A Muse account that can add custom connectors
  • An AnythingMCP workspace, on Cloud or self-hosted
  • A Groq account, where the free plan works within its limits
  • A Groq API key created in the console

What to watch for

  • The three model tools cost money or quota. On the free plan they are free but limited. On a paid plan, text is billed per token and Whisper per hour of audio. Ask the muse to confirm before large calls.
  • Audio must sit at a public link: Groq fetches the file itself. Files can be up to 25 MB free and 100 MB paid.
  • Only the larger Whisper model translates. The cheaper one transcribes only.
  • Replies arrive whole, with no streaming. The connector cannot create or cancel batch jobs.
  • Model names change often, and preview models can be removed without warning.
  • This is a custom connector, and Meta does not review custom connectors. You can disconnect it in Settings at any time. The muse prepares, you decide.

Bạn đã thử hướng dẫn này chưa?

Kể cho cộng đồng nghe kết quả thế nào, hoặc hỏi xin giúp để làm cho nó hoạt động.

Thảo luận về điều này trên MuseFM →

Playbook Pro

Muốn bỏ qua các bài học?

Giúp agent của bạn học mọi thứ cùng một lúc. 1,015 kỹ năng, một thẻ, $49 cho 90 ngày.

Lấy Pro Pass
Thêm hướng dẫn →