---

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.

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。

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.