Low-code studio

Design an agent in plain language, then inspect the Agno shape underneath.

Start with a template, describe the job in your own words, then shape the brain, tools, live connections, and specialist paths directly from this workspace.

Modes

Beginner first

Natural-language planning and plain words lead the experience, with Agno details shown alongside.

Runtime target

Agno on Cloud Run

Every hosted build is shaped around the DataFluid Agno runtime and an A2A-facing surface.

Publishing stance

Private by default

Builders can explore safely first, then choose public or unlisted visibility after testing tools and connections.

Starter templates

Pick a direction

Creator share: 70% of each successful charge. The rest stays with DataFluid for platform costs, payment operations, and support.

Selected template: Research scout. You can keep the starter prompt or rewrite it in your own words.

Choose the brain

How should this agent think?

Deep reasoner

Name and promise

Make the draft understandable

Users who need help gathering and comparing trustworthy information

Give it arms

Skills, MCPs, and specialist agents

1 selected
0 specialists0 live connections1 reusable skills

Saving turns on after you sign in and create your organization.

Draft overview

Research scout

Scans sources, compares options, and summarizes what matters in plain language. It starts with reference library.

Audience

Users who need help gathering and comparing trustworthy information

Who this first draft is designed to serve.

Brain

Deep reasoner

Useful when the agent must compare options, research carefully, or make harder decisions.

Price setup

$0.02 / 1K tokens

Start with a fixed credit price per run until you learn the average token cost of deeper reasoning jobs. Creator share: 70%.

Launch

private

This stays inside your workspace until you feel comfortable sharing it more broadly.

Unsaved draftGuest session
BrainModel | datafluid/deep-reasoner

Useful when the agent must compare options, research carefully, or make harder decisions.

Personality and rulesInstructions

Be calm, clear, and beginner-friendly in every answer. Ask a short clarifying question only when the next safe step is unclear.

ArmsTools, skills, MCP, live agents

Lets the agent answer from FAQs, policies, manuals, and company documents.

TeamworkTeams, routers, loops, and parallel work

Start with one main helper and only call specialists when the request is outside the safe default lane. Run source gathering in parallel when the user needs comparison work from multiple places at once.

Workflow preview

How the draft works

Main helperagent
Research scout

Handles the main conversation and keeps the experience friendly for non-technical users.

Do these at onceparallel
Parallel source lane

Lets the agent gather or compare multiple work paths together before responding.

Reference libraryskill
Reference library pack

Lets the agent answer from FAQs, policies, manuals, and company documents.

main-agentgather workparallel-lane
parallel-laneusesknowledge-library

Saved drafts

Workspace history

Draft history appears here after you sign in, create an organization, and save your first builder draft.

Capabilities

Recommended arms and specialists

Reference libraryDataFluid curated
Reference library pack

Lets the agent answer from FAQs, policies, manuals, and company documents.

Official or curated source | easy
4 starter brains7 starter capabilities6 editable templates

Agno preview

What this means underneath

Runtime model key

datafluid/deep-reasoner

Instructions

  • Act as Research scout.
  • Answer in plain language first and keep technical details optional.
  • State what you can do now, what needs a connected system, and when a human or specialist should step in.
  • Prefer the attached reference library before guessing.
  • Compare sources clearly and separate facts, unknowns, and recommendations.

Tools and connections

  • Reference library pack

Orchestration

  • Start with one main helper and only call specialists when the request is outside the safe default lane.
  • Run source gathering in parallel when the user needs comparison work from multiple places at once.

Launch advice

What to do next

  • Start private until your connections and guardrails behave the way you expect.
  • Use fixed price per call first, then refine token pricing once usage data is real.
  • Load your FAQs, policies, or operating notes early so the agent answers with your real language.