Providers & Models
Add your API key
Open Settings → Providers & keys, find the provider card, paste the key. It is encrypted on your machine. You can update keys at any time.
- Open the provider's console: console.groq.com/keys for Groq, platform.openai.com/api-keys for OpenAI, and so on.
- Create a new API key. Copy it once (most consoles don't show it again).
- Open Dimmy → Settings → Providers & keys.
- Click the card for your provider to expand it.
- Paste the key in the key field and click Connect.
- The card header shows a connected status pill when the key is accepted.
Where the key lives
- Encrypted at rest in ~/.config/dimmy/keys.enc (or %APPDATA%\dimmy\keys.enc on Windows).
- AES-256-CTR with an HMAC-SHA256 authentication tag, encrypt-then-MAC, so a tampered file is rejected instead of silently decrypted.
- The encryption key is derived on the spot from your username and your machine name. Nothing is written to an OS keyring, so there is no permission popup and no admin password on any platform.
- Decrypted only in memory when Dimmy needs to call the provider.
- Sent only to that provider's endpoint, over HTTPS. A URL with credentials in it, with line breaks, or on plain HTTP is refused, the one exception being a server on localhost.
Multiple keys
Each provider has its own card with its own key field. You can keep keys for Groq, OpenAI, Anthropic, and others simultaneously. Each card shows what the provider can do: speech to text, rewrite, recap.
One provider, up to three keys
Keys are stored per provider and per job, not one per provider. That means the same vendor can carry more than one account.
Most people never need the third one. It exists for the case where the recap should be billed to a different account than your day-to-day dictation, for example a work account for meetings and a personal one for everything else.
Was this helpful?
In this category
Providers & Models
Which model should you use?
Four questions decide it: what machine you have, what must not leave it, which subscriptions you already pay for, and which language you speak. Answer them and the rest follows.
Providers & Models
Local mode: transcription on your own machine
Three speech engines that run on your device. No API key, no upload, no per-minute cost. Works on a plane.
Providers & Models
Cloud providers compared
Groq is fastest, OpenAI is the gold standard, Deepgram leads on streaming. All optional. Bring your own key.
Providers & Models
Choose a local model
The whole on-device catalogue: nine Whisper sizes, Parakeet TDT v3, four Qwen3-ASR variants. What each one weighs and what it runs on.
Providers & Models
Optional on-device LLM
Run the LLM enhancement pass locally via llama.cpp. No cloud, no key, no per-token cost. Pick a model to match your GPU memory.
Last updated: 2026-05-22