Adding Voice Notes to Hermes Through Telegram
How I configured local speech-to-text with faster-whisper, vocabulary prompts, and selective Edge TTS audio replies for Hermes on Telegram.
find ./blogs -type f | sort --newest
Writing workspace / page 01
How I configured local speech-to-text with faster-whisper, vocabulary prompts, and selective Edge TTS audio replies for Hermes on Telegram.
Learn how to configure the Hermes Telegram gateway, set up user allowlists, manage background services, and leverage persistent AI memory on mobile.
A stateless chatbot is a helper. An always-on, self-hosted agent is an employee. This is the story of how I transitioned from localized terminal tools to setting up Hermes on a €5/month Hetzner VPS.
When Playwright introduced its new AI Agents, I couldn't resist trying them out on on a real webpage. In this blog, I'll walk you through how the Planner, Generator, and Healer with the support of Cursor helped me automate my Astro blog from scratch — what worked impressively well, what still needed a nudge from experience (and a few prompts), and what this says about the future of AI-assisted test automation.
In test automation, locators allow test scripts to interact with elements on a page, mimicking real user actions. Choosing the right locators can make tests more robust, reliable, and less prone to breaking with minor UI changes. Poor locator strategies can lead to flaky tests that fail for the wrong reasons, wasting time and resources. By choosing the right locator strategy you can create tests that are stable, easy to understand, and closely aligned with how users interact with your application, ultimately improving test effectiveness and maintainability. Playwright comes with a very wide and flexible way of locators, and in this blog post we will dive deep into how to define resilient locators and I will propose a locator strategy that I personally use on my projects.