
Daily Agents Need a Scheduler
The Mattermost Codex bridge moved recurring work out of immortal agent threads and into a small scheduler with state, receipts, and Mattermost …
// 20 posts in this category

The Mattermost Codex bridge moved recurring work out of immortal agent threads and into a small scheduler with state, receipts, and Mattermost …

A practical way to run Mattermost on a small server, expose it safely, and use the mobile app as a chat interface to a local Codex agent.

Supply-chain scanners should not bury local endpoint risk in raw JSON. Bumblebee has the engine; this UI turns the scan into something developers can …

AI coding agents do real repository work. They deserve review desks, not cramped chat boxes. This is the product shape behind a native Flutter client …

Dream TTS turned local speech from a model demo into product architecture: Flutter for the surface, Rust for durable synthesis work, and strict …

Local speech becomes trustworthy when pronunciation, latency, fixtures, logs, and release gates make quality repeatable instead of subjective.

A local speech feature survives only when the app talks to one stable product API while engines, voices, platforms, playback, and failures change …

Your first local voice ships only after it is bundled, checksummed, observable, cancellable, and boring enough to survive real user machines.

If your app needs speech without a hosted API, the hard part comes after the nice demo voice: engines, latency budgets, licenses, updates, and …
A postmortem on the Gravitee loopback pattern: the dead ends, the latency win, the sandbox boundaries, and the production checklist that matters next …
When a gateway cache only wants to cache full HTTP responses, the loopback pattern turns that constraint into a small internal token API.
What if a perfectly working API gateway was quietly fetching the same valid service token thousands of times a day? This field guide shows how to find …