Agents

An Agent wraps a backend with a role, instructions, memory and plugins.

from maki import MakiLLama
from maki.agents import Agent

llm = MakiLLama(model="llama3.2")
agent = Agent(
    name="Reviewer",
    maki_instance=llm,
    role="code reviewer",
    instructions="Focus on bugs, regressions, and missing validation.",
    stateful=True,
)

print(agent.execute_task("Review this design: a plugin system with file access."))

With stateful=True the agent keeps token-budgeted conversation history between tasks.

Memory and reasoning helpers

agent.remember("repo", "maki")
print(agent.recall("repo"))

steps = agent.think_step_by_step("How should we structure plugin validation?")
subtasks = agent.decompose_task("Prepare this repository for a public release")

Streaming

for chunk in agent.stream_task("Draft a short changelog entry."):
    print(chunk, end="", flush=True)

Long-running tasks

By default execute_task sends one blocking request. For tasks that may exceed the configured timeout (default 120 s), set use_streaming=True; the timeout then applies per chunk rather than to the whole response.

agent = Agent(name="Ranker", maki_instance=llm, role="content ranker", use_streaming=True)
result = agent.execute_task("Rank these 50 articles by relevance: ...")

Tools

Load a plugin and let the model call it:

agent.load_plugin("file_reader")
result = agent.execute_task("Read the first lines of README.md and summarize them.", use_plugins=True)

Verified tasks

A workflow task can carry a checker agent that must accept the result before the task counts as complete. See Workflows.