Skip to main content

Executors

AutoAgents ships with three primary executors. You can also implement your own by conforming to AgentExecutor.

  • Basic: single‑turn chat without tool calls
  • ReAct: multi‑turn reasoning with tool calls and streaming
  • CodeAct: multi‑turn reasoning where the model writes sandboxed TypeScript that composes tools

Basic​

Use BasicAgent<T> to run a single prompt/response cycle. Suitable for simple Q&A and when tools are not needed.

Key points:

  • No tool calls
  • Optional streaming variant
  • Output type: BasicAgentOutput → convert to your agent output via From<...>
  • Reasoning/thinking chunks are emitted in stream events (StreamChunk::ReasoningContent) when supported.
use autoagents::core::agent::prebuilt::executor::BasicAgent;
let agent = BasicAgent::new(MyAgent);

ReAct​

Use ReActAgent<T> for iterative reasoning + acting. Supports tool calls, multi‑turn loops (with max_turns), and streaming with intermediate events.

Key points:

  • Tool calls are serialized, executed, and their results fed back to the LLM
  • Emits events: tool requested/completed, turn started/completed, stream chunks
  • Output type: ReActAgentOutput → convert to your agent output via From<...>
  • Reasoning/thinking remains event-level (StreamChunk::ReasoningContent) and is not included in final output fields by default.
use autoagents::core::agent::prebuilt::executor::ReActAgent;
let agent = ReActAgent::new(MyAgent);

Tip: ReActAgentOutput::extract_agent_output<T> can deserialize a structured JSON response into your type when you expect strict JSON.

CodeAct​

Use CodeActAgent<T> when you want the model to compose tools through sandboxed TypeScript instead of issuing direct tool calls one by one.

Key points:

  • The model only sees one tool: execute_typescript
  • Your registered tools are exposed inside the sandbox as typed external_* functions
  • Each execution runs in a fresh native sandbox with explicit resource limits
  • Output type: CodeActAgentOutput { response, executions, done }
  • Emits additional events for code execution lifecycle and console output
  • Available on native targets with the codeact feature enabled
use autoagents::core::agent::prebuilt::executor::CodeActAgent;
let agent = CodeActAgent::new(MyAgent);