Skip to content

Agents

Agents are the core execution units in Synatra. Each agent is an AI executor with versioned configurations.

An agent consists of:

  • Working copy - Draft state for iterating on configurations
  • Releases - Published, immutable versions with semantic versioning
  • Model configuration - Provider, temperature, system prompt
  • Assigned tools - Functions the agent can call
  1. Navigate to Agents in the Console
  2. Click New Agent
  3. Configure the model settings
  4. Add tools as needed
  5. Write a system prompt

Synatra supports multiple LLM providers:

ProviderModelsSpecial Features
OpenAIGPT-4o, GPT-4 TurboReasoning effort control
AnthropicClaude 3.5 Sonnet, Claude 4Extended thinking
GoogleGemini 1.5 Pro, Gemini 2Thinking config

When you’re ready to deploy an agent:

  1. Review the working copy configuration
  2. Click Create Release
  3. The release becomes the active version

Releases are immutable and use semantic versioning (major.minor.patch). To make changes, update the working copy and create a new release.

Agents can generate dynamic UI widgets to interact with users:

Widget TypeDescription
FormInput forms with JSON Schema validation
TableSortable tables with row selection
ChartLine, bar, and pie charts
MarkdownFormatted text content
ConfirmationApprove/reject dialogs
Key-ValueDisplay structured data

Widgets can wait for user input before the agent continues, allowing for interactive approval flows and data collection.