Agent Playground

AI agent · Workflow / back-office agent · Operations & supply chain by Nango

Test AI agent integrations in a live dashboard sandbox.

A built-in AI agent sandbox within the Nango dashboard for testing and prototyping integrations across 1,200+ APIs. It provides developers with live observability and intent tracking to validate agent behavior before deployment.

Agent specs

The full spec card (context, pricing per token, modalities and more) on AntSpace. Sign up free to see the full spec card.

Launched
2026-10-20
Languages
English

Category: AI Agents

Channels
In-app (Nango Dashboard)
Deployment
Cloud (Nango Dashboard)
Setup
Minutes (connect apps and start asking tasks)
Supported APIs
1,200+
Pre-built Tools
10,000+
Interface
Nango Dashboard
  • Test AI agent behavior across 1,200+ APIs
  • Full transparency with intent tracking in logs
  • No-code environment for integration prototyping
  • Real-time observability of tool calls and latency

Overview

Agent Playground is a built-in AI agent environment within the Nango dashboard designed for developers and product teams to prototype integrations. It allows users to connect their own applications and test agent behavior across 1,200+ supported APIs before writing any production code. The playground provides full transparency by recording the user's goal and the agent's intent for every tool call, which can be inspected in real-time logs. Users can ask the agent to perform complex tasks, such as logging issues in Linear or updating deals in HubSpot, to validate how agents interact with external systems. It serves as a secure sandbox for refining integration logic and exploring real-world use cases through Nango's infrastructure.

Testing and Prototyping

  • Live Integration Testing: Users can connect their own applications and accounts to the playground to test integration behavior directly within the Nango dashboard.
  • Multi-API Task Execution: The agent can perform tasks across 1,200+ supported APIs and MCP servers, such as creating issues or updating records.
  • No-code Prototyping: Teams can explore and validate integration use cases before writing any production application code.
  • Sandbox Environment: The playground allows testing how an AI agent interacts with specific external systems through Nango's infrastructure.

Control and Observability

  • Intent Tracking: The dashboard records the user's specific goal and the agent's intent for every tool call made during a session.
  • Call Transparency: Users can open any call in the logs to see the reasoning behind the agent's specific actions.
  • Live Observability: The platform provides detailed traces of tool calls, including status, latency, and headers for debugging.
  • Connection Recovery: Agents can recover from broken connections or expired credentials without interrupting the active session.

What it does

  • Perform tasks across 1,200+ APIs\nLog issues in external tools\nUpdate CRM records\nTest integration logic end-to-end

Use cases

  • Developers connect their HubSpot and Linear accounts to test if an agent can correctly update a deal when a bug is logged.
  • Product teams explore new API integrations by asking the agent to perform tasks without writing any application code.
  • Engineers debug agent logic by reviewing the recorded intent and tool calls in the dashboard logs.

Specifications

Channels
In-app (Nango Dashboard)
Integrations
1,200+ APIs including Linear, HubSpot, Slack, Notion, and GitHub
Human control
Users provide goals and tasks; agent intent and tool calls are logged for human review and debugging
Deployment
Cloud (Nango Dashboard)
Setup
Minutes (connect apps and start asking tasks)
Supported APIs
1,200+
Pre-built Tools
10,000+
Interface
Nango Dashboard
Primary Use Case
Testing and Prototyping
Observability Format
OpenTelemetry (OTel)
Minimum Polling Frequency
30 seconds

How it works

  1. Access the Dashboard: Access the Agent Playground directly from the Nango dashboard.
  2. Connect Your Apps: Connect your external accounts (e.g., Salesforce, GitHub) using Nango's auth infrastructure.
  3. Define Tasks: Provide a goal or task to the built-in AI agent, such as updating a deal or creating a ticket.
  4. Observe and Refine: Review the agent's intent, tool calls, and logs to validate the integration logic.

FAQ

What is the Nango Agent Playground?

The Agent Playground is a built-in environment in the Nango dashboard where you can connect your apps and ask an AI agent to perform tasks directly before writing code.

Which APIs can I test in the playground?

It supports 1,200+ APIs and 10,000+ pre-built tools, including popular platforms like Linear, HubSpot, and Slack.

What happens if a connection breaks during a session?

If a connection's credentials expire or lack a scope, the user receives a reconnect link to authorize again, allowing the agent to continue the session without interruption.

Explore use cases

Added by AntSpace from public sources (nango.dev).

More AI agents by Nango