AI News Feed
Market watch
LG UltraFine Evo 32U990A-S Hits Record-Low $1,099 on Amazon, TechRadar ReportsNewegg Labor Day Deal: 32GB Kingston Fury Beast DDR4-3200 Kit Nearly Half Price at $179.99 with CodeGalaxy Z Fold 8 inner screen corners reported with excessive giveHuawei unveils Mate XT 2 tri-fold with privacy-screen edition as Apple prepares folding iPhoneWall Street Analysts Recommend Nvidia, Uber and Marvell for Long-Term GrowthLenovo, AMD and Huawei Put Local Large-Model AI at Center of New Devices
LG UltraFine Evo 32U990A-S Hits Record-Low $1,099 on Amazon, TechRadar ReportsNewegg Labor Day Deal: 32GB Kingston Fury Beast DDR4-3200 Kit Nearly Half Price at $179.99 with CodeGalaxy Z Fold 8 inner screen corners reported with excessive giveHuawei unveils Mate XT 2 tri-fold with privacy-screen edition as Apple prepares folding iPhoneWall Street Analysts Recommend Nvidia, Uber and Marvell for Long-Term GrowthLenovo, AMD and Huawei Put Local Large-Model AI at Center of New Devices
Companies

Chaitin Technology Launches MonkeyCode AI Coding Platform with Built-in Security Scanning

Chaitin Technology has launched MonkeyCode, an open-source AI coding platform with a built-in security scanning engine. It supports private deployment and multiple Chinese LLMs, and offers a free tier of 30 million tokens per day.

AI News Feed · 2026-08-17 22:53

The company says it has more than ten years of experience in security attack and defense research, and it built MonkeyScan to perform real-time security audits on AI-generated code. The engine detects common vulnerabilities including SQL injection, cross-site scripting, command injection, and insecure encryption algorithms. According to Chaitin, the scanning covers OWASP Top 10 and CWE Top 25 risk categories and includes specialized rule sets for Java, Python, JavaScript, and Go. The company also says MonkeyScan is deeply coupled with code generation rather than being an external post-processing tool.

MonkeyCode’s source code is publicly available on GitHub, where the project has more than 3,600 stars, according to the company. Chaitin says the AGPL-3.0 license lets users audit how the platform handles prompts, data transmission, and storage, which it argues is not possible with closed-source competitors. The platform can be deployed entirely inside an organization’s intranet using Docker, and all AI inference can run locally. It supports NVIDIA GPUs as well as domestic Huawei Ascend accelerators. The company also highlights a browser-based environment that requires no software installation and native iOS and Android apps for mobile code review.

For model support, MonkeyCode has adapted to a set of Chinese LLMs including GLM, Kimi, MiniMax, Qwen, and DeepSeek. The company says a multi-model routing system selects the best model for each task based on the coding language, prompt context length, and historical success rates. This design, it says, prevents vendor lock-in and allows models to be swapped without interrupting development workflows.

Pricing information released by the company shows a free plan with 30 million tokens per day, which it says is enough for typical personal use. Paid plans are priced at ¥99 per month for the professional tier, ¥499 per month for the flagship tier, and ¥299 per user per month for the enterprise tier. In its materials, Chaitin compares this with Cursor’s $20 monthly subscription and Claude Code’s estimated daily token costs of $6 to $13, arguing that MonkeyCode’s free allowance is more than adequate for most individual developers.

In a separate technical whitepaper, Chaitin says corporate demand for AI development tools has shifted from simple code completion to requirements such as security auditability, model flexibility, data boundary control, and teamwork consistency. It describes MonkeyCode’s security architecture as four layers: prompt injection defense, MonkeyScan code analysis, dependency and license compliance checks, and full audit logging. The whitepaper also positions MonkeyCode against platforms such as Cursor, GitHub Copilot, OpenAI Codex, Trae, WorkBuddy, and Qoder, citing differences in licensing, pricing, and deployment options. One point it emphasizes is that closed-source tools cannot be independently audited.

Chaitin says its private deployment solution has been validated in production environments at multiple financial institutions, though it does not name them. It also says the platform has been designed to comply with strict data-residency requirements in sectors such as finance, government, and healthcare.