Dictio does not have a transcription server. When you hold the hotkey and speak, your audio goes from your device straight to OpenAI, using an API key that belongs to you, and the text comes straight back to the field you were typing in. That is the whole path. Nothing of yours passes through our servers, because there are none in the way.
What an API key is, and where to get one
An API key is a long string that tells OpenAI which account to bill for a request. You create one in your OpenAI account under API keys, copy it once, and paste it into Dictio’s settings. Dictio stores it on your device and uses it only to send your speech for transcription. You can revoke the key from your OpenAI account at any time, and Dictio simply stops working until you paste a new one.
What we never see
- Your audio. It travels from your device to OpenAI and nowhere else.
- Your text. The transcript comes back to your device and is typed where your cursor is. We do not receive a copy.
- You. There is no Dictio account to create, no telemetry, and no training on your voice.
What it means for your bill
You pay OpenAI directly for the minutes of audio you send, at the rates on their pricing page, and you can watch usage in your own dashboard. There is no middleman margin and no monthly plan needed to start dictating. Dictio itself is free to start.
Why we built it this way
Dictation is intimate. You will say things into it that you would never type into a web form: drafts, half thoughts, names, numbers. The only honest way to promise that your voice stays yours is to never have it in the first place. Bring-your-own-key is not a technical shortcut. It is the point.
