Human In The Loop Python Packages

swarmkit-runtime

Open-source runtime for governed multi-agent AI — define agents, tools and human-approval gates as data (YAML), run under real gates, compiled to LangGraph.

osprey-framework

An agentic interface and safety harness for safety-critical control systems

squidbrake

Brakes for your AI agents: every tool call is checked against your rules, held for human approval when risky, and recorded in a tamper-evident audit trail. Works with Claude Code and any MCP app.

ctrlrun

The execution safety layer for AI agents.

briefloop

本地多运行时 Agent 工作台：围绕材料研究、生成可编辑报告、独立复核、追溯证据，并从人工反馈持续学习。

noeta-sdk

Python SDK for durable AI agents: survive crashes, pause for human approval at no cost, scale from a script to a multi-host cluster. Any LLM (Anthropic, OpenAI-compatible), MCP, self-hosted.

swarmkit-schema

Open-source runtime for governed multi-agent AI — define agents, tools and human-approval gates as data (YAML), run under real gates, compiled to LangGraph.

privacyfence

Open-source privacy and approval gateway between AI assistants and your business systems (MCP): human approval, local PII checks and audit for Gmail, Drive, Slack, Salesforce, Jira and more.

langstage-core

Host + AG-UI bridge for LangGraph agents — agent-spec loading, layered config, and typed in-process streaming

agenda-intelligence-md

Deterministic evidence-packet checks for AI-generated claims: references, quotes, numbers, and support gaps. CLI, Python, MCP, and review adapters.

autoagent-core

Zero-dependency LLM agent runtime — tool use, MCP, sandboxed self-written tools, checkpoint/resume, approval gates, bi-temporal fact memory, code-level guards (context bounds, loop guard, egress control, shadow mode), declarative tool policy, OpenTelemetry. Not a framework. 885 tests, CI Linux/Windows, 31 runnable demos, online visual builder.

nestor-meaning

Has a human checked this? Seal/serve/audit over verified-match memory — ledger, MCP, and CLI.

noeta-runtime

Python SDK for durable AI agents: survive crashes, pause for human approval at no cost, scale from a script to a multi-host cluster. Any LLM (Anthropic, OpenAI-compatible), MCP, self-hosted.

shyftlabs-continuum

Continuum — the agent runtime by ShyftLabs. Build, orchestrate, ship.

langstage-cli

The terminal stage for your LangGraph agent — a Claude Code-style CLI for any CompiledGraph

agentraas

The reliability layer for what AI agents do: exactly-once execution, action policies, human approval and a full audit trail. Self-hosted, MCP-native, free Community edition (MIT/Apache-2.0).

pushary-openai-agents

Phone approvals for OpenAI Agents SDK agents. Your agent asks, your user taps Approve or Deny. TypeScript and Python.

leashd

Safety-first agentic coding framework. Three-layer safety pipeline (sandbox, YAML policies, human-in-the-loop approval) for AI coding agents. Pluggable runtimes (Claude Code, Codex), autonomous task orchestrator, full audit trail.

vibey-engine

Open-source orchestrator for AI coding agents: a spec interview, unattended builds on local (Ollama) or paid agents, human-gated review, and an append-only PostgreSQL ledger so a crashed agent's work is picked up, not lost.

yuanzhu

Meta-Forge Workshop: turn recurring work into governed AI workflow assets. MCP server + Claude Code plugin + dsh plugin — staged writes, evals-gated forging, inference engine. 元铸工坊 SOSOFAST：把最佳实践铸成可治理的 AI 工作流资产

pushary

The decision layer for AI agents. Your agent asks, your user decides on their phone. Python SDK.

proton-safe-mcp

Draft-only Proton Mail MCP server with attachment support and out-of-band human approval.

boundflow-charter

Build and operate production-safe agents that run in your own environment.

xero-trial-balance-export

Runs deterministic month-end close checks on Xero trial-balance exports and produces exception packs for human sign-off.