Find the distinctive concepts hiding in your code.
Concept Scanner reads your codebase and surfaces the distinctive technical concepts inside it. It runs entirely on your hardware using a local LLM, so your code never leaves your machine. No cloud uploads, no API keys, no external calls.
Code scanners ask what is wrong. Concept Scanner asks what is distinctive.
Not a patent scanner. It reads your code for distinctive technical concepts. That is engineering work, not legal work. Whether any concept is worth patenting is a separate question for a patent attorney.
Why local matters for patents.
Unfiled inventions are trade secrets. Uploading your code to cloud AI services before filing can compromise patent rights and create foreign filing license complications. From scan to structured disclosure in minutes. Minimize your exposure window.
Complete confidentiality
Your source code and analysis results stay on your machine. No data is transmitted to any external service. Critical for pre-filing confidentiality.
Works offline
After the initial model download, the scanner operates entirely offline. No internet connection required. No API keys to manage or pay for.
Filing compliance
Patent documents processed on foreign servers can create foreign filing license issues under 35 U.S.C. 184. Local processing avoids this entirely.
What the scanner finds.
The scanner produces structured technical disclosures, not vague suggestions. Every discovery links back to the specific code that contains the innovation.
Structured concepts
Each discovery includes a technical description and problem-solution framing, structured for review by your team or an attorney.
Grounded, with context
Every concept traces back to specific files and code sections, with a validation state, plus git authorship and dependency-license context. Verifiable provenance a reviewer can check against the actual implementation.
Scored and classified
Each mechanism is scored on four engineering axes (technical distinctiveness, implementation depth, problem specificity, generality) and classified as distinctive, borderline, or textbook.
Nothing hidden
Textbook concepts are held back by default, but the count is always reported. The scanner never silently drops a finding.
Three steps to your first scan.
Download and install
Grab the binary for your platform. Install Ollama for the local LLM runtime. One-time setup takes about 5 minutes. On first run, a memory-aware picker detects your RAM and recommends a model profile that fits; a larger model surfaces more.
Point at your codebase
Run the scanner against any local directory. It reads your source files, analyzes patterns, and identifies concepts that may be distinctive and worth protecting.
./concept-scanner scan ./your-project --verbose
Review discoveries
Examine scored strategic concepts with evidence from your actual code. Each discovery links to specific files and lines so you can verify the analysis.
Download.
All binaries are published on GitHub. Free and open source.
Requirements.
Ollama
Local LLM runtime. Free and open source. Install from ollama.ai
8 GB RAM minimum
16 GB recommended for larger codebases and faster analysis.
4 GB disk space
For the LLM model download. One-time download on first run.
Your code. Your machine. Your patents.
Download Concept ScannerNot sure which fits? Compare all the ways to run the scanner.