Edict vs Spokenly.
Spokenly takes the AI-agent side of dictation seriously, which makes this the closest comparison on the site. Their wins first, and there are more of them than you might expect.
Spokenly is a dictation app with a free tier, iOS support, and a documented MCP server; it can use local or cloud speech models. Edict is a $29 one-time macOS app with no cloud path at all: models ship inside the app and dictation audio cannot leave the Mac.
Both apps ship a built-in MCP server for AI agents. Edict’s serves transcripts, dictionary editing, and on-device file transcription to agents like Claude Code.
Last verified 2026-07-25
Where Spokenly wins
Their free tier is two tiers, both described as $0 forever: local models with unlimited use that work completely offline, and a bring-your-own-keys tier they charge nothing for. Unlimited offline dictation costs nothing here, which is more than most rivals give away and more than Edict does. There are iOS and Windows apps, which Edict doesn’t have. And they ship a real, documented MCP server so agents can request voice input, with support for local models so speech can stay on-device.
Their comparison’s best-pick table calls Spokenly the "only one with a built-in MCP server". Read in context, that line is scoped to the four apps the comparison covered, and it was fair within that scope. It just isn’t true of the market: Edict ships one too, and this page exists partly so the record is checkable.
The architectural difference
Spokenly offers a choice of speech engines, local and cloud. Flexibility again, and again a legitimate feature. Edict’s position is narrower on purpose: the models are in the box, speech recognition runs on-device on the Neural Engine, and there is no mode, setting, or provider by which dictation audio can leave the Mac. To be precise about the difference, because it is narrower than it sounds: Spokenly ships a documented Local Only Mode that blocks network access and, in their words, "Guarantees that no data leaves your device". That is a real guarantee. The distinction is only that theirs is a switch you can turn off and ours is an absence of code, which matters if you worry about misconfiguration and not otherwise.
The second difference is philosophy about your words. Edict is verbatim by design: what you said, plus your own dictionary’s spelling fixes, never an AI rewrite. Your prompt to an agent should be your prompt.
Side by side
Competitor cells verified against Spokenly’s site and docs on the stamped date.
| Spokenly | Edict | |
|---|---|---|
| Price | Two $0-forever tiers; Pro $99.99 a year | $29 one-time |
| Platforms | macOS, iOS, Windows, and Linux | macOS only (15+, Apple Silicon) |
| Speech processing | Choice of local or cloud engines, with a Local Only Mode that blocks the network | On-device only; no cloud path exists |
| Models | Selected and downloaded in-app | Bundled inside the app; offline from first launch |
| MCP server | Yes, documented | Yes: transcripts, dictionary, file transcription |
| Verbatim guarantee | Optional AI processing available | Always verbatim; dictionary fixes only, never a rewrite |
| Auto-send to agents | Voice input via MCP request flow | Push dictation: per-app allowlist, signal on the pill |
| File transcription | Yes, in-app and CLI with SRT/VTT export; diarization provider-dependent | Diarized SRT / VTT / text, about 150x real time |
How to choose
Pick Spokenly if you want a free tier, iOS, or engine flexibility, and you’re comfortable choosing which requests stay local.
Pick Edict if you want one price once, a hard offline guarantee with the models already in the box, and verbatim text into whatever you’re working in. If your day is Claude Code or Cursor, read the walkthrough and judge by the workflow, not this table.
Common questions
Is Spokenly the only dictation app with a built-in MCP server?
No. That claim, where it appears, is scoped to the four apps in their own comparison. Edict ships a built-in MCP server too: agents can search transcripts, pull the latest one, edit the dictionary, and submit files for on-device transcription.
Is Spokenly free?
It has a free tier with paid upgrades; check their pricing page for current details. Edict is a one-time purchase with a 7-day fully functional trial.
Which keeps speech more private?
Spokenly can run local models, so speech can stay on-device depending on configuration. Edict removes the choice: there is no cloud path in the app, so dictation audio cannot leave the Mac under any setting.
Does either app rewrite what you say?
Edict never does: verbatim plus your own dictionary fixes is the contract. Spokenly offers optional AI processing of transcripts, which some users want.
Apple Silicon (M1 or newer) · macOS 15 Sequoia or later · 7-day full trial · one-time purchase