Quickstart

Connect an AI provider

Open LLM providers in the sidebar (or jump straight in via Studio later). Paste a key from your provider's dashboard. The key is encrypted at rest and never leaves your organization.

You can connect more than one provider and give each agent the model that suits it — a fast cheap one for a simple classifier, a strong one for an agent that writes careful answers.

More on AI providers →

Create your first agent

Every agent belongs to an Application. Open your Application's Agents page and click New Agent. The dialog asks for:

• Name — Research Briefing
• Application — the Application the agent belongs to. With only one Application, it's picked for you; with several, it starts on the one you're in.

Click Create agent. You land in Studio, where you configure the agent.

Configure the agent

An agent is one form: what it takes in, what it's told to do, and which model does the work. Fill in two groups:

1. Instructions
   • Parameters — click Add parameter:
     • Name: topic
     • Type: STRING
     • Required: ✓
   • System Prompt: You are a research analyst. Write concise, factual briefings.
   • Input Prompt:

     Write a one-page briefing on: {{param.topic}}
     
     Include: background, key players, recent developments, why it matters.
     Use short paragraphs. No fluff.

2. Model & capabilities — under LLM Config, pick the provider you connected, then a Model.

Insert {{param.topic}} from the prompt editor's placeholder menu. On each run, Fruxon fills it with the run's topic, and the model's reply is what the agent returns. Studio saves your edits as you go.

Test it

Use Test chat on the right. Pick Single run, fill topic with something — "the EU AI Act" — and run.

You'll see:

• The model's response, streamed live
• A full trace showing prompts, tokens, latency, cost
• Any error context if something went wrong

Iterate freely. Your edits stay in a draft until you deploy, so nothing you try reaches production.

Deploy

Happy with it? Click Deploy. Say in a line why you're making the change; Fruxon saves the draft as a new version (a revision), runs its pre-deploy checks, and makes it live. The deployed revision is what API calls, triggers, and participant-routed messages will hit. To roll back, deploy a previous revision — there's no downtime.

More on versioning and rollback →

Call it from anywhere

Your agent is now reachable three ways:

• REST API — POST /v1/{tenant}/agents/{agent}:execute with a Fruxon bearer token. See the API reference.
• Python SDK — pip install fruxon. See the SDK guide.
• Agent Network — Bind Slack, Teams, Telegram, Twilio, or other provider channels to participants, and bind those participants or triggers to the agent. See Agent Network.