Add AI Dictation as an open-source Superwhisper alternative

Following [CONTRIBUTING_TO_OSB.md](https://github.com/junaid33/opensource.builders/blob/master/CONTRIBUTING_TO_OSB.md), this is a submission for an open-source alternative to the existing Superwhisper record.

Add open source tool: AI Dictation
- Repository: https://github.com/writingmate/aidictation
- Website: https://aidictation.com
- Description: Open-source voice-to-text for macOS, Windows, iPhone, iPad, and Android, with offline speech recognition on supported devices and optional cloud transcription and text cleanup.
- Category: Productivity & Note-taking
- Suggested proprietary baseline: Superwhisper
- Features: Desktop global-shortcut dictation, an iOS voice keyboard, an Android floating microphone, offline recognition on supported devices, optional cloud transcription and cleanup, and personal vocabulary or writing controls where supported by the platform.

## Additional information

- License: MIT
- Platforms: macOS, Windows, iOS/iPadOS, and Android
- Self-hosting: Not claimed. This repository provides native client applications, not a documented self-hosted service.
- Offline caveat: Offline recognition and a fully offline workflow are different. Optional cloud cleanup may send transcript text when enabled, and feature availability varies by platform.
- Suggested existing capabilities for maintainer review: `speech-to-text`; `global-hotkey` and `auto-pasting` for desktop; `offline-first` only with the caveat above; and `ai-text-transformation` for optional cleanup.

Disclosure: I am affiliated with AI Dictation. I used AI assistance to prepare this submission and manually checked it against the current app, website, repository, and contribution guide. I have not assigned a similarity score because that editorial judgment belongs to the catalog maintainers.

Start with CONTRIBUTING_TO_OSB.md and inspect the existing Superwhisper record to follow the catalog's entry format. Add AI Dictation with the supplied repository, website, description, category, license, platforms, and carefully qualified offline and cloud-transcription details; done means the listing is accurate and its suggested capabilities are suitable for maintainer review.

Written by the indexing model from the issue text.