Unvibe complex code. Trace business workflows.
Complex code hides connections, workflows, and business risks. Unvibe it.
Explore connected code in an interactive graph, give LLMs the right context, trace business workflows, and identify critical business risks.
UnvibeCode is useful when an important business workflow is distributed across multiple files and reviewers need more than isolated file summaries.
| Audience | Their concern | What UnvibeCode gives them |
|---|---|---|
| Product managers | Missing states, edge cases, requirements, and failure paths | Business workflow maps showing decisions, state changes, material effects, and scenarios that require validation |
| Developers | Understanding where a workflow exists across unfamiliar or complex code | Connected entry points, functions, calls, dependencies, effects, and downloadable LLM-ready context |
| Engineering leaders | Prioritizing critical code paths and understanding the consequences of change | Connected workflow paths, affected components, supporting evidence, and risk-focused review priorities |
| Founders | Revenue, customer, operational, permission, and data-integrity risks | Business-risk summaries explaining the trigger, current behaviour, affected workflow, and potential impact |
| QA and test engineers | Determining what should be validated beyond the happy path | Evidence-backed failure paths and structured scenarios that can be converted into test cases |
| LLM users | Giving an LLM enough connected repository context without sending unrelated files | Complete repository context, smaller connected-code packages, and verified relationships |
Use UnvibeCode when you need to:
- Understand an unfamiliar, inherited, legacy, or LLM-generated codebase.
- Trace a business operation that crosses routes, services, state, databases, queues, and external APIs.
- Give an LLM connected repository context without manually selecting files.
- Review registration, payment, booking, subscription, approval, notification, or other multi-step workflows.
- Investigate failure paths involving retries, partial completion, authorization, concurrency, or external dependencies.
- Understand the potential business consequence of a code-level problem.
- Prepare for a significant change to a business-critical workflow.
- Produce a reusable repository-context package for later LLM analysis.
pip install unvibecode
python -m unvibecode review --repository "/path/to/repository"Windows example:
python -m unvibecode review --repository "D:\upstox"No activation key. No OpenAI API key. The repository path is the only required input.
UnvibeCode displays live progress, opens the completed reports, and saves the results automatically under ./shipready_results.
The Connected Code Map for LLMs shows how files, symbols, imports, and calls work together.
Hover over a file to preview it. Click the file to trace its connected code and download the relevant code as one compact, LLM-ready package.
Stop manually selecting files or pasting unrelated code into an LLM.
The Complete Repository Context organizes code into reusable chunks with filenames, line references, verified connections, and contextual selections. Use the complete ZIP later, or download a smaller connected-code package directly from the graph.
complete_repository_context_for_llm.zip
├── manifest.json
├── chunks.jsonl
├── connections.jsonl
├── selections.jsonl
└── README.md
A file tree explains repository structure, but it does not explain how the software performs business operations.
The Business Workflow Map connects entry points, decisions, state changes, external calls, and business outcomes into understandable workflows.
The Business Risk Findings report identifies critical workflow risks in the code and shows the supporting evidence.
Each finding explains:
- What triggers the risk
- What the code currently does
- The affected business workflow
- The potential business impact
- What should change
- How to verify the correction
- The exact supporting code evidence
If no finding passes the evidence threshold, the report clearly records a no-findings outcome for the completed review scope.
UnvibeCode scans supported source files and resolves meaningful file, symbol, import, and call relationships.
Code is organized into reusable chunks and connected packages. Selecting a file retrieves the surrounding code needed to understand it without sending the entire repository to an LLM.
Related code paths are interpreted as workflows containing triggers, decisions, state changes, external effects, and business outcomes.
Completed workflows are checked for failures that may affect customers, payments, permissions, data integrity, operations, or other important business outcomes.
For a supported repository, the automatically created results folder contains:
shipready_results/
└── analysis_<timestamp>/
└── customer_results/
├── 01_connected_code_map_for_llm.html
├── complete_repository_context_for_llm.zip
├── 02_business_workflow_map.html
└── 03_business_risk_findings.html
UnvibeCode supports code understanding, workflow review, and evidence-based engineering investigation.
It does not replace:
- Unit, integration, end-to-end, load, or penetration testing.
- Manual code review by engineers familiar with the system.
- Security, privacy, legal, or regulatory assessment.
- Production monitoring and incident investigation.
- Validation of business requirements with the responsible product owner.
A reported scenario should be validated against the application's intended behaviour and runtime environment before being treated as a confirmed production defect.
Supported repositories receive the complete code, workflow, and risk review.
Repositories above approximately two million estimated source tokens receive the Connected Code Map and Complete Repository Context. The deeper Business Workflow and Risk Review does not run above that threshold.
Read How UnvibeCode works for details.
Repository parsing, dependency mapping, connected-context preparation, and report rendering run locally.
Structured context required for business workflow and risk analysis is securely sent to the hosted UnvibeCode service. Provider credentials remain server-side.
Review only repositories you are authorized to process.
Read the complete data-processing explanation.
- Python 3.11 or newer
- Internet access for hosted business analysis
- Read access to the repository being reviewed
For installation problems, failed reviews, report-interpretation questions, public-repository review requests, collaboration enquiries, or general feedback:
- Open a GitHub issue for reproducible package problems.
- Email divya.singaravelu@iiml.org for product questions, public-repository review requests, or collaboration enquiries.
- Report security vulnerabilities through the process described in SECURITY.md.
When requesting technical support, include:
- UnvibeCode version
- Python version
- Operating system
- Repository language and framework
- Command used
- Sanitized error message
- The report stage that failed
Do not email proprietary source code, credentials, access tokens, secrets, or unsanitized logs.
Maintaining or evaluating a public repository?
Email divya.singaravelu@iiml.org with:
- The public GitHub repository URL
- The workflow or code area you want to understand
- The question you want the analysis to answer
- Permission to publish the resulting maps or findings, if applicable
Only submit repositories that you are authorized to analyse.
- Quick start for Windows, macOS, and Linux
- Understanding the four outputs
- How UnvibeCode works
- Data processing and privacy
- Troubleshooting
- Public preview policy
- Version history
- Read the contribution guidelines before proposing a change.
- Report vulnerabilities through the process described in SECURITY.md.
- Reuse and distribution are governed by the repository's license.