chat-model-configuration

name	chat-model-configuration
description	Configures agent behaviour via @agent() parameters and AgentConfig. Use when adjusting max_turns, temperature, max_tokens_per_turn, max_cost_usd, tool_error_policy, memory window, or when the agent must stop after a fixed number of turns.

Use codemap find "SymbolName" to locate any symbol before reading — it gives exact file + line range and is faster than grep across the whole repo.

Chat Model Configuration & Parameter Tuning

Quick start

# agent.py — from __future__ import annotations is safe here (no @tool)
from __future__ import annotations

from lauren_ai import agent
from lauren_ai._memory._stores import InMemoryConversationStore

@agent(
    model="claude-opus-4-6",
    system="You are a helpful assistant.",
    max_turns=3,
    temperature=0.7,
    max_tokens_per_turn=2048,
    memory_window_tokens=20_000,
    tool_error_policy="return_error",
    conversation_store=InMemoryConversationStore(),
)
class ConfiguredAgent: ...

---

@agent() parameters

Parameter	Type	Default	Description
model	str	required	Model identifier, e.g. "claude-opus-4-6"
system	str	""	System prompt for every turn
max_turns	int	10	Loop iteration cap; raises AgentMaxTurnsError
max_tokens_per_turn	int	4096	Output token limit per LLM call
temperature	float	1.0	Sampling temperature (0.0–2.0)
memory_window_tokens	int	40_000	Sliding context window in tokens
max_cost_usd	float | None	None	Cost cap; raises AgentBudgetExceededError
parallel_tool_calls	bool	False	Run all tool calls in a turn concurrently
tool_error_policy	str	"return_error"	"raise" | "return_error" | "skip"
conversation_store	ConversationStore | None	None	Persistent conversation history
memory	ShortTermMemory | None	None	Per-agent shared short-term memory

---

max_turns enforcement

from lauren_ai import agent, AgentMaxTurnsError
from lauren_ai._agents._runner import AgentRunnerBase as AgentRunner
from lauren_ai._config import LLMConfig
from lauren_ai._transport._mock import MockTransport

@agent(model="claude-opus-4-6", max_turns=2)
class LimitedAgent: ...

cfg = LLMConfig(provider="anthropic", model="mock-model", api_key="mock")
mock = MockTransport()
runner = AgentRunner(transport=mock)

# Queue 5 tool-use completions — runner stops at max_turns
# AgentMaxTurnsError is raised once the cap is hit

---

Lifecycle-doc class (using a docstring as system prompt)

@agent(model="claude-opus-4-6", max_turns=5)
class DocAgent:
    """You are an expert Python programmer.
    Always respond with working code.
    """

The class docstring becomes the system prompt when no system= kwarg is given.

---

Reference files

File	Contents
src/lauren_ai/_config.py	AgentConfig frozen dataclass — all defaults
src/lauren_ai/_agents/__init__.py	@agent() decorator, AgentMeta
src/lauren_ai/_agents/_runner.py	AgentRunnerBase loop with max_turns check