Palash Dev
Palash Debnath is an independent software engineer and product builder behind palash.dev, building mostly open-source products around AI voice, local-first media, boot media, and persistent memory for coding agents. His stated aim is to build things he wishes existed and help change the global agentic-software industry from Northeast India.
At a Glance
- Independent creators and media producers
- Developers and AI-tool builders
- Teams needing private/local voice and media workflows
- Coding-agent users and software engineers
- +1 more
AI Tools by Palash Dev
(1)
VoiceStudio
Local Voice AI Desktop App
Discussions
No discussions yet
Be the first to start a discussion about Palash Dev
Latest News
VoiceStudio v0.5.3 released: new look and Electron desktop implementation
Opal v0.8.5 released with startup, playback, Windows and omnibox improvements
Bootable v0.1.4-rc.pr8.1 released with cross-platform packages and safety controls
memXT v0.4.0 released with prebuilt local binaries
Products & Services
Open-source AGPL-3.0 desktop studio and local API for voice cloning, voice design, video dubbing, dictation, transcription, stories and audiobook creation. It supports local engines on CPU, CUDA, Apple Silicon/MLX, AMD ROCm or Docker, and offers an OpenAI-compatible local API plus MCP connectivity.
Cross-platform native desktop app and mouse-enabled terminal UI for finding, writing and verifying ISO, IMG, RAW and compressed images to USB or SD media. It lists removable media, excludes system disks, revalidates targets and verifies hashes and written bytes.
MIT-licensed, local-first long-term memory for Claude Code and coding agents: a single Zig binary with MCP tools and session hooks, local embeddings, llama.cpp and sqlite-vec, with no API key or network required at query time.
GPL-3.0 native media player/browser for macOS, Linux and Windows that combines search, discovery, chat, browsing and playback. It accepts titles, local files, magnet links, servers, feeds or natural-language moods and uses a Zig core with source plugins.
Market Position
Palash positions VoiceStudio as an open-source, fully local alternative to cloud voice platforms such as ElevenLabs: it emphasizes no API key, no subscription or usage meter for local work, local processing, 646 languages, swappable engines, a desktop app, an OpenAI-compatible API and MCP support. Opal is positioned against fragmented combinations of media players, browsers and catalog apps, while memXT targets cloud-dependent or heavier AI-memory tools with a small local binary and on-device search.
Leadership
Founders
Palash Debnath
Independent software engineer and product builder; GitHub profile describes him as a builder from Northeast India who builds mostly open-source software. His public work spans VoiceStudio, Opal, Bootable, memXT and other developer tools.
Founding Story
Palash describes palash.dev as his personal corner of the internet and his work as building products he wishes existed, mostly open source. The recurring product thesis is local-first ownership: keep voice, media state, agent memory and other sensitive workflows on the user's machine rather than renting them from cloud services. His VoiceStudio origin story is specifically a reaction to voice cloning and dubbing being available mainly through metered APIs that require sending a user's voice to a third party.
Business Model
Revenue Model
The local VoiceStudio application is free and open source under AGPL-3.0, with no account, subscription or usage meter for core local workflows. The VoiceStudio site also describes a commercial Pro offering and a hosted Cloud offering in early access; the public Cloud API is not released.
Pricing Tiers
Open-source desktop app; local workflows require no account or API key and have no usage meter.
Commercial-rights and hosted-access offering described on the VoiceStudio website.
Hosted product; the public API is described as a preview contract and has no production endpoint advertised.
Target Markets
- Independent creators and media producers
- Developers and AI-tool builders
- Teams needing private/local voice and media workflows
- Coding-agent users and software engineers
- Desktop users on macOS, Linux or Windows
- Creators producing voiceovers, dubs, audiobooks and narrated stories
- Developers integrating local speech through an OpenAI-compatible API or MCP clients such as Claude and Cursor
- Teams needing commercial or hosted VoiceStudio access (Pro is enquiry-only and Cloud is in early access)
- Coding agents that need persistent, searchable project context across sessions
- Users who want a single native media discovery and playback surface
- Users creating verified bootable USB or SD media