Adapt
Mini toolsCassini 1.0
By LatentForce
Cassini-1.0 is a LatentForce small language model fine-tuned for static code feature extraction. Given a source file, it outputs structured JSON metadata for dependency resolution, call graph construction, type inference, impact analysis, and large-codebase intelligence workflows. It is trained for Python, JavaScript, and TypeScript and extracts six symbolic fields in one pass: imports, definitions, references, calls, type assignments, and rich definition metadata. The model is based on Qwen3-4B, uses supervised fine-tuning with LoRA, supports up to 20,480 tokens, and is released under Apache 2.0.
New
Text
Gen 7
Overview
Cassini-1.0 is LatentForce’s 4B code-intelligence model for extracting structured symbolic metadata from Python, JavaScript, and TypeScript source files.
About LatentForce
LatentForce is an AI-native platform for large-scale enterprise code migration and software modernisation. It uses agentic AI workflows and task-specific Small Language Models to transform legacy codebases into modern architectures with deterministic, audit-ready migration pipelines, reducing migration cost, risk, and time by up to 80%.
View Company ProfileTools using Cassini 1.0
No tools found for this model yet.
Last updated: June 9, 2026
