Liangpei: AI Matchmaking and Emotional Companion App, Focusing on Precise Matching and Compliant Operations
Founded: Zeng Xinxun · Liangpei Technology
Key Fields
FIELD STAMPSOrigin
Zeng Xinxun previously served as the product lead for Moonshot AI's Kimi team. Observing that traditional matchmaking apps had a fake account proportion exceeding 15% and a matching accuracy of less than 10%, a large number of single users faced dual pain points of matchmaking needs and emotional companionship. After large model technology matured in 2024, AI could solve core issues such as identity verification, personalized matching, and emotional chat, prompting Zeng to leave and start a company to build a compliant AI matchmaking and companionship product.
Milestones
Turning Points
- In June 2024, Zeng Xinxun left the Kimi team and officially entered the AI matchmaking and companionship track.
- In August 2025, completed Series A financing, moving from the product validation stage to the scaled expansion stage.
- In February 2026, the AI simulated dating feature went live, completing the closed-loop between virtual companionship and real-world matchmaking.
Failures & Pitfalls
- The AI virtual lover feature launched in January 2025 was removed from the app store due to compliance risks.
- The offline matchmaker + AI matching pilot model in November 2024 was suspended because single-event manual costs exceeded the budget by 2x.
- In April 2026, targeted by malicious complaints from traditional matchmaking platforms, the app was demoted by the app store for 3 days, causing downloads to drop by 40%.
关键成功要素
- The core team originates from Moonshot AI's Kimi team, possessing deep accumulations in large model NLP technology.
- Public security identity verification + liveness detection fundamentally resolves the pain point of fake accounts on matchmaking platforms.
- Compliant operations in vertical matchmaking scenarios avoid content safety risks.
- The dual-demand superposition of emotional companionship and matchmaking enhances user willingness to pay and retention.
Lessons
- Emotional companionship products must place compliance first to avoid red-line risks such as pornography and fraud.
- Precise matching in vertical tracks has a higher lifetime value than users in pan-social scenarios.
- Teams with a technical background starting a business need to supplement operations and commercialization capabilities as early as possible.
- Do not blindly expand into non-core scenarios; focusing on the core pain points of target users is essential to boost conversion rates.
Core Data
- 月经常性收入:6 million RMB (Public data basis, independent review not verified)
- 累计注册用户:4.5 million (Public data basis, independent review not verified)
- 付费用户数:123,000 (Public data basis, independent review not verified)
- 种子轮融资额:10 million RMB (Public data basis, independent review not verified)
- 团队规模:65 people (Public data basis, independent review not verified)
- 虚假账号占比:0.3% (Public data basis, independent review not verified)
- 首轮融资额:50 million RMB (Public data basis, independent review not verified)
Competitors / Peers
Liangpei's peers and benchmarks include traditional matchmaking platforms like世纪佳缘 and 珍爱网, as well as AI emotional companionship products like Replika and 星野. Traditional platforms have massive user bases but generally suffer from a fake account proportion exceeding 15%, matching accuracy below 10%, and high user complaint rates. Pure AI emotional companionship products have strong user willingness to pay but lack matchmaking landing scenarios, failing to meet users' rigid demand for marriage. Relying on AI identity verification and precise matching models, Liangpei boasts a fake account proportion of only 0.3%, and its paid conversion rate is 2.3 times that of pure companionship products, presenting clear differentiated advantages.