Your first parallel workflow in five minutes

Parallel orchestration sounds complex until you run it once. This walkthrough gets you from a blank account to a finished multi-agent answer in about five minutes — no custom agents or API keys beyond your model provider.

1. Create an account and add a key

Sign up at masternode.ai and open Settings → Model providers. Paste an API key for at least one supported provider (OpenAI, Google, Anthropic, Groq, DeepSeek, or Cohere). MasterNode never resells inference — you pay the provider directly and keep full control of spend.

2. Pick a prompt that benefits from parallelism

Single-sentence questions work, but you'll feel the difference with tasks that naturally split: a competitive brief, a release note from several source files, or a research summary with citations. Paste the goal into chat and send.

• Planning runs first — you'll see a task graph before agents execute.
• Independent branches run at the same time instead of waiting in a queue.
• The merged answer streams back as sections complete.

3. Read the live status panel

Each subtask shows as running, complete, or blocked on a dependency. If something needs web grounding, search runs only when the intent classifier says it's required — not on every message.

4. Iterate

Follow up in the same thread. Keyword memory keeps durable facts (your preferred units, product names, tone) without stuffing the full transcript back into context. Attach PDFs or spreadsheets when you need answers grounded in your own documents.