Getting Started

Pitchblend is a private desktop workspace for chatting with AI models on your own computer. You do not need to understand model formats, tokens, or inference engines before you begin. This section explains those terms when they first matter.

Recommended path

1. Read What Pitchblend is for a plain-language overview and an honest explanation of what "local" means.
2. Follow Installation and system requirements for your operating system.
3. Use First launch and model setup to get the local runtime ready.
4. Complete Your first conversation.
5. Take the Interface tour so you know where Chat, Workflows, Files, Models, Settings, account, and Memory Sync live.

If local AI is new to you

Start with Local AI basics. It explains models, downloads, context, memory, RAG, tokens, and why the first answer can take longer than later answers.

A safe first goal

For the simplest first run, install the curated model offered on the Models > Local Models tab, make it Active, and ask a short text question in Chat. Add files, voice, online providers, workflows, account sign-in, and cross-device memory sync only after ordinary chat works.