---
title: ConnectOnion Screening Card
company_id: connectonion
company_name: ConnectOnion
dd_case_id: a3ac978f-0dfe-4eff-a6c3-d4eaff6e6786
canonical_document: screening_card
created_at: '2026-03-23T20:39:24+00:00'
last_updated_at: '2026-03-23T20:39:24+00:00'
update_history:
- '2026-03-23T20:39:24+00:00 :: Initial creation of Screening Card based on preliminary
  risk assessment and market research.'
---

> Canonical DD Screening Card
> Case ID: `a3ac978f-0dfe-4eff-a6c3-d4eaff6e6786`
> Last updated: 2026-03-23T20:39:24+00:00
> Update note: Initial creation of Screening Card based on preliminary risk assessment and market research.

<!-- KOROK:LANG=zh -->
# 项目初筛卡: ConnectOnion

**日期:** 2026-03-23
**分析师:** AI Due Diligence Agent
**阶段:** Pre-seed / Seed
**赛道:** AI 基础设施 / 智能体框架 (AI Agent Frameworks)

## 一句话描述
一个极简的 P2P AI 智能体框架，专注于极致的开发者体验 (DX) 和无缝的跨组织智能体协作。

## 评分表

| 评估维度 | 分数 | 信号 | 理由 |
| -------- | ---- | ---- | ---- |
| 团队 (Team) | 6/10 | 🟡 | 创始人 Aaron Xie 技术能力强且具有韧性，但 2 年内经历 3 个项目，长期专注度待观察。 |
| 市场 (Market) | 6/10 | 🟡 | 智能体编排赛道需求旺盛，但正面临协议标准化的快速整合期。 |
| 竞争 (Competition) | 4/10 | 🔴 | 面对 LangChain (LangGraph) 和 CrewAI 等巨头，以及 Anthropic MCP 协议的生态挤压。 |
| 产品/技术 (Product/Tech) | 7/10 | 🟢 | “10倍快”的开发者体验和内置可观测性 (@xray) 是核心亮点，但 P2P 发现机制存在安全隐患。 |
| 财务/表现 (Financials) | 5/10 | 🟡 | 获得 UNSW 10x 加速器 $100k 资助；PyPI 下载量 2.8万+，GitHub Star 真实增长。 |
| 风险 (Risk) | 7/10 | 🔴 | 核心风险在于协议孤岛化和 P2P 网络所需的临界规模效应。 |
| **总分 (Overall)** | **5.8/10** | **🟡** | **高潜力挑战者，风险与机遇并存。** |

评分标准: 🟢 >= 7 | 🟡 4-6 | 🔴 <= 3
风险分值反向: 🟢 <= 3 (低风险) | 🔴 >= 7 (高风险)

## 核心优势 (Top 3)
1. **极致的开发者体验 (DX)**：开发者反馈上手速度比 LangChain 快 10 倍，Python 原生感强，有效解决了现有框架过度抽象的痛点。
2. **差异化的 P2P 协作模型**：避开了主流的中心化编排，通过 P2P Relay 实现跨组织智能体发现，具有独特的网络效应潜力。
3. **内置可观测性 (@xray)**：产品工具链完整，自带调试与链路追踪，降低了智能体开发和维护的门槛。

## 核心疑虑 (Top 3)
1. **协议标准化冲击 (MCP 竞争)**：随着 Anthropic MCP 成为行业事实标准，ConnectOnion 的私有 P2P 协议可能面临沦为“孤岛”的风险。
2. **P2P 网络安全与信任**：去中心化发现机制在缺乏身份验证 (DID) 的情况下，易受“工具投毒”和供应链攻击。
3. **创始人的战略定力**：创始人频繁转型的历史 (ApplyOnion 等) 让投资人对其能否在 3-5 年的长期协议建设中保持专注存疑。

## 结论
**建议:** 深入跟进 (Deeper Look)
**理由:** 尽管面临巨头竞争和标准化压力，ConnectOnion 通过极佳的 DX 成功切入了开发者群体。目前的 2.8 万次下载量证明了其产品的真实吸引力。

## 创始人访谈重点问题
1. **协议兼容性**：ConnectOnion 计划如何实现与 Anthropic MCP 和 Google A2A 的深度兼容？
2. **安全机制**：目前的 P2P Relay 机制如何防止恶意智能体或工具注入？是否考虑引入基于 DID 的零信任验证？
3. **长期愿景**：考虑到过去两年的多次转型，创始人如何确信 ConnectOnion 是值得投入 5-10 年的长期赛道？
4. **增长策略**：在悉尼/UNSW 社区之外，如何在全球范围内（如硅谷）获取开发者心智？

> ⚠️ 本尽调报告基于公开信息和 AI 分析生成。应作为人类判断、背景调查和直接创始人互动的补充，而非替代。
<!-- /KOROK:LANG -->

<!-- KOROK:LANG=en -->
# Deal Screening Card: ConnectOnion

**Date:** 2026-03-23
**Analyst:** AI Due Diligence Agent
**Stage:** Pre-seed / Seed
**Sector:** AI Infrastructure / Agent Frameworks

## One-Liner
A minimalist P2P AI agent framework focused on extreme developer experience (DX) and seamless cross-organizational agent collaboration.

## Scorecard

| Dimension    | Score | Signal   | Justification |
| ------------ | ----- | -------- | ------------- |
| Team         | 6/10  | 🟡 | Founder Aaron Xie is technical and resilient, but has cycled through 3 projects in 2 years; long-term focus needs monitoring. |
| Market       | 6/10  | 🟡 | High demand for agent orchestration, but the sector is undergoing rapid protocol consolidation. |
| Competition  | 4/10  | 🔴 | Faces intense pressure from giants like LangChain (LangGraph), CrewAI, and the Anthropic MCP ecosystem. |
| Product/Tech | 7/10  | 🟢 | "10x faster" DX and built-in observability (@xray) are key; however, P2P discovery has security vulnerabilities. |
| Financials   | 5/10  | 🟡 | Secured $100k AUD from UNSW 10x; 28k+ PyPI downloads with organic GitHub growth. |
| Risk         | 7/10  | 🔴 | Existential risks include protocol isolation and the need for critical mass in the P2P network. |
| **Overall**  | **5.8/10** | **🟡** | **A high-potential challenger with significant execution and standardization hurdles.** |

Scoring: 🟢 >= 7 | 🟡 4-6 | 🔴 <= 3
Risk score inverted: 🟢 <= 3 (low risk) | 🔴 >= 7 (high risk)

## Key Strengths (Top 3)
1. **Extreme Developer Experience (DX)**: Feedback suggests onboarding is 10x faster than LangChain, with a Pythonic feel that solves the "abstraction soup" problem.
2. **Differentiated P2P Collaboration Model**: Eschews centralized orchestration in favor of a P2P Relay for cross-org agent discovery, offering unique network effect potential.
3. **Built-in Observability (@xray)**: Comes with integrated debugging and tracing tools, lowering the barrier to entry for developing and maintaining complex agents.

## Key Concerns (Top 3)
1. **Protocol Standardization Impact (MCP Competition)**: As Anthropic’s MCP becomes a de facto industry standard, ConnectOnion’s proprietary P2P protocol risks becoming an isolated "walled garden."
2. **P2P Security & Trust**: The decentralized discovery mechanism currently lacks robust Zero Trust verification (e.g., DIDs), making it vulnerable to tool poisoning and supply chain attacks.
3. **Founder Longevity & Focus**: A history of frequent pivots (e.g., ApplyOnion) raises concerns about the founder's ability to stay the course for a multi-year protocol play.

## Verdict
**Recommendation:** Deeper Look
**Reasoning:** Despite fierce competition and standardization pressures, ConnectOnion has captured developer mindshare through superior DX. 28k+ downloads indicate genuine product traction.

## Questions for Founder Meeting
1. **Protocol Compatibility**: How does ConnectOnion plan to achieve deep compatibility with Anthropic’s MCP and Google’s A2A?
2. **Security Roadmap**: How does the P2P Relay prevent malicious agents/tools? Is there a plan for DID-based verification?
3. **Long-term Vision**: Given the pivots over the last two years, what makes you certain that ConnectOnion is the long-term horse to bet on?
4. **Global Growth**: Beyond the Sydney/UNSW hub, how do you plan to scale the developer community globally, especially in Silicon Valley?

> ⚠️ This due diligence report is generated from publicly available information and AI analysis. It should supplement, not replace, human judgment, reference calls, and direct founder interactions.
<!-- /KOROK:LANG -->