HIGHCOURT expands legal AI knowledge base and review tools

HIGHCOURT said its private legal AI platform now includes 31.3 million indexed legal passages and improved review performance through MKC2. The update is aimed at law firms that want AI support for research and drafting while keeping processing on their own infrastructure.

Why it matters: - HIGHCOURT is targeting law firms that want AI-assisted legal research and review without moving core processing off their own infrastructure. - The expanded corpus and review model are meant to help attorneys sort evidence, assess authorities and prepare case materials faster. - The company is positioning the platform as a private legal AI workspace for U.S. legal matters.

What happened: - HIGHCOURT announced expanded legal knowledge and specialized review capabilities for its private legal AI platform on October 4, 2026. - The platform is already available for deployment. - HIGHCOURT said the update improves legal reference coverage and specialized review functions. - The platform supports legal research, matter preparation, drafting and review in a connected workspace.

The details: - HIGHCOURT’s installed legal reference infrastructure includes 31.3 million indexed legal passages across case law, statutes, regulations and supporting materials. - The system includes 9.98 million case records in its citation database. - The platform tracks 77.5 million citation relationships to show connections between legal authorities. - Coverage includes federal law and regulations, all 50 states and additional legal reference resources. - Coverage depth varies by jurisdiction and practice area. - Passage counts reflect indexed text segments, including overlapping collections, not unique cases or individual laws. - MKC2, also called HC Critic, reviews generated work and flags potential problems for human review. - In an internal evaluation, MKC2 reached 93.17% accuracy on scored binary checks, compared with 83.09% for the comparison model. - MKC2 posted a 0.962 ROC-AUC on ranking defective work above sound work, compared with 0.728 for the comparison model. - The performance figures come from the company’s internal evaluation set and do not represent live-matter outcomes or a guarantee of error-free legal work. - MKC2 was developed by HIGHCOURT founder Michael Knight, who also founded Privilege AI.

Between the lines: - HIGHCOURT is emphasizing measurable scale and review quality, two features that matter most when legal teams want AI assistance but still need human oversight. - The company is also drawing a line between internal test results and real-case performance, which signals caution around overreading benchmark claims. - The retrospective case validations suggest the platform is being used in real legal workflows, but the sample is small and unpaid. - A participating law firm reported no hallucinations or issues in the three case outputs it reviewed.

What's next: - HIGHCOURT said the platform is available now for deployment. - Product information and deployment inquiries are available at HIGHCOURT’s website. - The company’s corpus figures reflect installed indexes checked October 4, 2026, and the MKC2 results reflect a September 21, 2026 internal calibration record.

The bottom line: - HIGHCOURT is betting that larger legal coverage plus stronger automated review can make private, attorney-controlled AI more useful for law firms.