Getting started
From install to your first shipped project in a few steps.
1. Install the app
Download PromptConnext for macOS, Windows, or Linux from the download page and open it. No account is required to start.
2. Connect a model
On first run, onboarding guides you to connect and health-check at least one working model. You have two options:
API key / endpoint — paste a key or an OpenAI-compatible base URL (OpenAI, Anthropic, Google, OpenRouter, and more).
Local, zero cost — install Ollama, pull a model such as qwen3:8b, and select “Local Ollama.”
3. Create a project
Start a new project and describe what you want to build in plain language. This is the Scope stage — PromptConnext generates a structured specification for you.
4. Review the spec, then build
Approve the generated plan in the Spec stage. Then, in Skill, connect a coding model or agent (Claude Code, Gemini CLI, or a custom agent) and let implementation proceed — every step traceable in the task graph.
Your code and model keys never leave your machine. Only the shared task graph syncs to the cloud, and only when you choose to collaborate.
Start building with the AI you already have
Download PromptConnext, connect a model in minutes — including a zero-cost local option — and take your first project from scope to running code.