managed-research 1.2026.42111

Public SDK, MCP, and schemas for Managed Research.

managed-research

Managed Research lets you start hosted AI workers from Python or MCP and inspect their work as durable runs. It is for repo and research tasks where you want logs, checkpoints, artifacts, approvals, usage, and final outputs instead of an untraceable chat transcript.

Install

uv add managed-research

export SYNTH_API_KEY="sk_..."

Quickstart

import os

from managed_research import ManagedResearchClient

client = ManagedResearchClient(api_key=os.environ["SYNTH_API_KEY"])

run = client.runs.start(
    "Review the project context and propose the smallest high-impact improvement.",
    host_kind="daytona",
    work_mode="directed_effort",
    providers=[{"provider": "openrouter"}],
)

print("run:", run.run_id)

result = run.wait(timeout=60 * 60, poll_interval=15)
print("state:", result.state.value)
print("artifacts:", [artifact.title for artifact in run.artifacts()])

ManagedResearchClient is the canonical entrypoint. SmrControlClient remains available as a compatibility alias for one release.

Main Ideas

• Use client.runs.start(...) for a one-off run on the default project.
• Use client.projects.create(...) and client.project(project_id) for durable project-scoped work.
• Attach repositories, files, datasets, credentials, notes, and knowledge before starting a run.
• Use preflight when you want launch blockers as structured data before spending runtime.
• Inspect runs through messages, timeline, traces, checkpoints, artifacts, usage, questions, approvals, and actor/task counts.
• Use agent_harness="codex" or agent_harness="opencode_sdk" when you want to pin the harness explicitly.

OpenCode Harness

OpenCode is a first-class harness option with this initial model palette:

• anthropic/claude-sonnet-4-6
• anthropic/claude-haiku-4-5-20251001
• x-ai/grok-4.1-fast

For deeper examples, see:

• docs/quickstart.md
• docs/python-sdk.md
• docs/mcp.md