examples/

Command activity_pipeline_example mirrors Python's examples/activity_pipeline_example.py — the local-first activity observation pipeline (Fases 0-4).

Command agent_pool_example mirrors Python's examples/agent_pool_example.py — run three specialized agents in parallel, each on a different task, via multi.Pool.

Command basic_agent mirrors Python's examples/basic_agent.py — the smallest possible Simon agent: build one with defaults and run a prompt.

Command builtin_tools_agent mirrors Python's examples/builtin_tools_agent.py — running an agent's built-in tools directly via the "tool:name {json_args}" shorthand.

Command chat_tui mirrors Python's examples/chat_tui.py — an interactive terminal chat with a named, personality-driven agent.

Command hooks_agent mirrors Python's examples/hooks_agent.py — observability hooks and usage tracking via agent.WithOnEvent.

Command knowledge_agent mirrors Python's examples/knowledge_agent.py — index a PDF into the knowledge base, then ask the agent questions that can only be answered from that document.

Command knowledge_router_agent demonstrates Knowledge Router: a hierarchical, lexical, embeddings-free alternative to Simon's vector KnowledgeBase.

Command mcp_agent mirrors Python's examples/mcp_agent.py — using tools from an MCP server inside a Simon agent.

Command memory_agent mirrors Python's examples/memory_agent.py — demonstrates conversation memory across two sequential Run calls.

Command parallel_agents mirrors Python's examples/parallel_agents.py — run three specialized agents in parallel over the same prompt via multi.Group.RunAll.

Command persistent_memory_agent mirrors Python's examples/persistent_memory_agent.py — one JSON file == one conversation.

Command planner_agent mirrors Python's examples/planner_agent.py — decompose a goal into tasks and run each one.

Command public_basic_agent is the smallest possible consumer of the public simon SDK: build a Runtime, open a Session, run a prompt.

Command public_cancellation demonstrates Session.Cancel: a slow scripted Model is interrupted mid-flight, and the resulting event stream ends with run.cancelled instead of run.completed.

Command public_desktop_wails shows the event-forwarding pattern a desktop application built with Wails (https://wails.io) would use to pipe simon.Event values into its frontend.

Command public_knowledge demonstrates attaching a knowledge base to a Runtime and surfacing a retrieved hit through a Session.Run call, using a scripted Model so the run is deterministic and needs no embedding API.

Command public_memory demonstrates attaching persistent memory to a Session via a MemoryFactory, and shows history surviving across multiple Run calls on the same session.

Command public_parallel_sessions demonstrates running multiple Sessions on one Runtime concurrently, and WithMaxConcurrentRuns throttling how many of those runs execute at once.

Command public_streaming demonstrates Session.Stream: consuming the <-chan simon.Event as a run progresses instead of waiting for the final Response.

Command public_structured_output demonstrates simon.RunStructured: a scripted Model replies with raw JSON (inside markdown fences, to show that fences are stripped), which RunStructured parses into a typed Go struct.

Command public_tool_approval demonstrates ApprovalPolicy: a custom policy denies a "delete_file" tool call and allows everything else, showing both outcomes without ever touching the filesystem.

Command public_tools demonstrates registering a typed tool and driving a full tool-call round trip: a scripted Model requests the tool on its first reply, then produces a final answer once given the tool's result.

Command run_context_example is an idiomatic Go adaptation of Python's examples/run_context_example.py, NOT a literal port.

Command structured_output_agent mirrors Python's examples/structured_output_agent.py — structured output parsed into a typed Recipe struct via agent.RunStructured.

Command tool_runner_example mirrors Python's examples/tool_runner_example.py — tool.Runner, Simon's standalone, turn-by-turn tool-use loop.

Command triage_agent mirrors Python's examples/triage_agent.py — a triage agent routes tasks to the right specialist via multi.NewTriage.