Latest 'Pay-as-you-go AI' Including Claude Mythos/Fable 5 Analyzes Emails and Adds New Features to mailSpeak to Instruct Alexa and Apps

IMD Co., Ltd. (Headquarters: Minato-ku, Tokyo; President & CEO: Junichi Mineo) has added a new feature to its mailSpeak service that makes it easy to use AI triggered by emails, integrated into the incoming mail filter.

Flow of AI Analysis and Notification Forwarding

The available AI models include the latest 'Claude Fable 5' and 'Claude Mythos 5', as well as Gemini, GPT, and Claude, characterized by the flexibility for users to specify prompts.

Delivering AI Decisions to the Real World Triggered by Emails

The content of incoming emails and user-specified prompts are sent to the designated AI via API. Based on the response (analysis results) and the specification/judgment of notification destinations, mailSpeak delivers the information to Alexa, email addresses, or apps.

Support for Gemini/GPT/Claude

Users can input API keys from the three AI companies and use them on a pay-as-you-go basis according to their needs and costs. Since users simply input the API keys of the AI services they have contracted with and use them at cost, the service itself is provided essentially for free.

Specifying Processing with Prompts

Freely Specify Prompts

Users can give detailed instructions on processing policies, enabling operations that go beyond standard summarization.

A key strength is the ability to specify abstract expressions such as 'if it's an important email' or 'only when judged to be urgent'. As a result of these expressions, users can safely select specific actions and destinations like 'send email', 'app notification', or 'notify via Alexa'.

AI Specifies Notification Destination

A 'command' system has been prepared so that the AI can judge the necessity and dynamically select the forwarding destination (Alexa, email address, app). mailSpeak recognizes the AI's response and acts accordingly.

To prevent misrecognition, the response begins with a #mailSpeak tag, followed by commands that can specify settings including the email forwarding destination, app, and app volume. These instructions are generated from the prompt, and a 'system prompt input support' function has been prepared to automatically input easily specified prompts.

To track what actions were taken, an 'AI analysis execution log' for prompts and responses to the AI is also provided. (By default, logs are saved only up to a certain length to prioritize privacy, but with permission, all logs can be saved for several days for operational verification.)

Further Flexible Support for Advanced Users

System prompt input support, pre-specifiable response results, and detailed response content settings are possible.

Specify App Read-Aloud Volume

For app forwarding, the AI can even specify the read-aloud volume, allowing it to be used for various notifications, such as changing the volume based on importance.

Bringing Digital Information to the Real World

The mechanism for determining the notification destination from emails and prompts was essential for mailSpeak as a service that connects the digital space to the real space.

It converts changes in the digital space understood by the AI into forms that users in the real space can notice, such as Alexa's read-aloud, smartphone app notifications, and email forwarding.

Enhanced Features for Free and Paid Versions

The free version has also been upgraded with features like forwarding for charging and incoming calls. The premium version now supports up to 10 managed devices, improving its capacity for heavy use.

Service Comparison

Reference Values for API Usage

Consideration for API Key Usage

Generally, API keys cannot be used within the scope of AI companies' subscription plans; users need to purchase additional API keys. Paying after use carries the risk of not seeing usage limits.

Therefore, mailSpeak recommends prepaid contracts with limits, and for AIs that provide clear token usage information, it allows users to check their usage at any time. Notifications for exceeding a certain usage amount are also planned.

Screen Images

This mailSpeak settings screen can be accessed from the app by tapping 'My Page' and then 'Mail Filter and Alexa/Webhook Settings'. From the web, log in at https://mailSpeak.jp/.

Tap each image to enlarge it.

mailSpeak Web Screen, Incoming Mail Log, Mail Filter Settings, Mail Filter Items

URLs/Links

mailSpeak - https://mailSpeak.jp/

mailSpeak App - https://mailSpeak.jp/app/

X - https://x.com/mailspeak_

Instagram - https://www.instagram.com/mailspeak_/

IMD Co., Ltd. - https://imd.jp/

Release Materials - https://mailspeak.jp/release/release-20260611.html

Android App (Google Play)

mailSpeak is available in the following forms:

- Android App

- Web App

- Alexa Skill

The Alexa Skill is only required for Alexa integration.

The Android App and Web App are linked to each other.

Accounts are shared, so Google ID authentication is recommended. (There are currently no plans to support iOS, which cannot handle notifications.)

FAQ

What is mailSpeak? It is a service operated and developed by IMD Co., Ltd., founded in 1996.

It started as an Alexa skill that reads new emails aloud, and has evolved into an AI-friendly service that converts digital information into the real space, including app-based smartphone notification forwarding.

What can the new features do?

New AI-related features have been added.

The AI understands the email content and sends the analysis results to a specified destination.

The AI understands the email content and can dynamically determine processing and notification destinations according to prompt instructions.

For emails classified by the mail filter settings, AI analysis is sent to a specified email address, Alexa, or app only for important emails or urgent messages.

Only when the AI judges that forwarding is necessary due to urgency, it responds with a 'command' on the first line, which mailSpeak recognizes and acts upon. Currently, it uses commands, but future support for MCP is planned.

This service is built for users who want to give detailed instructions on AI-based 'email processing' according to their intentions.

While there are many services for summarizing or responding to emails with AI, this service features a 'command' function that notifies the real world without limiting processing results to the digital space, and even allows the notification destination to be dynamically specified via prompts. AI models can also be selected based on cost-effectiveness.