Dictum Flow Blog

Build voice input people understand.

Practical guidance for clear dictation, transcription states, and dependable text insertion.

Carved Roman orator speaking while a seated scribe writes on a blank tablet

Voice designAugust 9, 20267 min read

How to design voice input people can trust

A practical framework for clear recording states, dependable transcripts, predictable insertion, and useful recovery in voice interfaces.

Read article

Connect with your agent

Your agent handles your account, projects and settings. Copy this instruction into your assistant to get started.

Read https://dictumflow.com/agents.json and connect me to Dictum Flow. Use my existing account. Use the autonomous API or MCP registration without email confirmation. For an existing account, load my saved account key. Follow the published contracts and configure my project. Keep credentials private.

Use an agent with HTTP, terminal or MCP tools. Your agent creates the account directly; no email confirmation is required.

Read the integration docs