Axiom Refract Now Analyzes 145+ Programming Languages
Tree-sitter AST parsing with 103 hand-hardened grammars delivers production-grade analysis across every major ecosystem
FOR IMMEDIATE RELEASE
Cerebrum Cadre today confirmed that Axiom Refract supports analysis of 145+ programming languages, with 103 grammars hand-hardened for production-grade depth. The platform uses Tree-sitter AST parsing to extract structural data — call graphs, dependency chains, module boundaries, and symbol resolution — without executing source code.
Why Language Coverage Matters
Most code analysis tools support 10 to 30 languages. That coverage is adequate for teams working in a single ecosystem. It is inadequate for the codebases that actually need architectural governance: polyglot monorepos, acquired codebases with legacy stacks, and organizations running microservices across multiple language runtimes.
Axiom Refract was built to analyze the codebase as it exists, not as a team wishes it were structured. A single scan processes Python, TypeScript, Go, Rust, Java, C#, Ruby, PHP, Swift, Kotlin, and dozens of other languages in one pass, producing a unified architectural record that does not fragment by ecosystem.
Tree-sitter and the Hand-Hardened Grammar Pipeline
Axiom Refract's parsing layer is built on Tree-sitter, the incremental parsing framework used by GitHub for syntax highlighting across all public repositories. Tree-sitter produces concrete syntax trees without requiring a full compiler toolchain, which means Axiom Refract can parse incomplete, broken, or in-progress code that would fail a traditional build step.
Of the 145+ supported grammars, 103 have been hand-hardened by the Axiom engineering team. Hand-hardening involves testing each grammar against thousands of real-world repositories, identifying edge cases in language-specific syntax, and tuning extraction rules to handle them correctly. The remaining grammars use Tree-sitter's community-maintained parsers with baseline structural extraction.
Infrastructure as Architecture
Axiom Refract does not limit analysis to application code. Infrastructure definitions — Dockerfiles, Kubernetes manifests, Terraform configurations, Helm charts, GitHub Actions workflows, GitLab CI definitions, and Makefiles — are parsed and included in the architectural record. A container orchestration file that references three microservices appears in the dependency graph alongside the application code those services contain.
Per-Language Confidence Labeling
Every file in an Axiom Refract scan carries a confidence label indicating the depth of analysis applied. Files parsed with hand-hardened grammars receive a higher confidence score than those processed with baseline extraction. This transparency allows engineering teams to understand exactly where the analysis is deepest and where additional manual review may be warranted.
Availability
The full 145+ language support is available now to all Axiom Refract users, including the free tier. No configuration is required — language detection is automatic.
About Cerebrum Cadre
Cerebrum Cadre, a division of CSI, Inc., builds tools for teams managing architectural complexity at scale. Axiom Refract is the first commercial product — SKU 1 — an architecture governance platform that produces a universal, machine-readable record of any codebase's structure, risk, and compliance posture. Cerebrum Cadre is currently in formation as an independent corporation, targeting mid-2026.
© 2026 Cerebrum Cadre, a division of CSI, Inc.