Dictum SDK · browser

Desktop and Mobile Web Test Console

SDK pending idle

Transcription timing

Monotonic offsets in milliseconds. Client and Worker clocks have separate origins.

Raw timing data

No timing result.

Effective model: —
Audio duration: —
Start to recording: —
Quota reservation: —
Audio ingress span: —
Transcript span: —
Stop to result: —
First transcript − last audio: —

Two-form replacement test

Each form owns its textarea. Starting the other form replaces the active page session.

Form 1 selected client ready

Ready

Dictation inactive.

Volume 0 %

Browser scenario

Mandatory initial transport: WebSocket, with automatic POST fallback.

Alan · signed identity demo

Ready to ask the simulated customer backend to sign Alan.

1. Alan claimspending
2. Customer backend signaturepending
3. Dictum Worker validationpending microphone start

Advanced: paste another customer JWT

Identity signed by the customer backend

The token is never stored, placed in the URL, or displayed in logs.

State: idle
Public operation ID: none (by design)
Retryable: false
Identity: required
Initial transport: websocket
Final transport: pending
Fallback: none

Browser compatibility

Support: unknown
Capture: unknown
Missing: none

Manual typing remains available without voice control.

Public Worker Contract

Endpoint
Project
Parity: checking…
Request: none
States: 0 / 7
Events: 0 / 18
SDK errors: 19
Transports: websocket · post
Names: distinct · 16 characters max.

Unverified local JWT claims

Display-only decoding. Only the Worker verifies the signature and authorizes each transcription.

kid: unknown
name: unknown
sub: unknown
aud: unknown
iat: unknown
exp: unknown

Last safe SDK error