# Docs

> Agentic Reverse Engineering

## Docs

- [Interpreting Kong Output](https://docs.kong.fyi/case-studies/interpreting-output.md): How to read and understand Kong's analysis results
- [Other Backdoors](https://docs.kong.fyi/case-studies/other-backdoors.md): More real-world case studies coming soon
- [XZ Backdoor (CVE-2024-3094)](https://docs.kong.fyi/case-studies/xz-backdoor.md): How Kong autonomously reconstructed the XZ Utils supply chain attack in 15 minutes
- [Call-Graph Analysis](https://docs.kong.fyi/concepts/call-graph-analysis.md): How Kong uses bottom-up call graph ordering to give the LLM better context for every function it analyzes.
- [Context Windows](https://docs.kong.fyi/concepts/context-windows.md): What goes into each LLM prompt — and why raw decompiler output is not enough.
- [Deobfuscation](https://docs.kong.fyi/concepts/deobfuscation.md): How Kong identifies and removes obfuscation techniques using an LLM-driven symbolic tool loop
- [Pipeline Overview](https://docs.kong.fyi/concepts/pipeline-overview.md): Kong's five-phase analysis pipeline: triage, analysis, cleanup, synthesis, and export — how each phase works and why the ordering matters.
- [Semantic Synthesis](https://docs.kong.fyi/concepts/semantic-synthesis.md): How Kong unifies naming conventions, renames globals, and synthesizes structs in a single post-analysis LLM pass.
- [Signature Matching](https://docs.kong.fyi/concepts/signature-matching.md): How Kong identifies known library functions before LLM analysis, saving time and money
- [Syntactic Normalization](https://docs.kong.fyi/concepts/syntactic-normalization.md): How Kong cleans up Ghidra's decompiler output before sending it to the LLM
- [Type Recovery](https://docs.kong.fyi/concepts/type-recovery.md): How Kong recovers struct definitions from pointer access patterns across multiple functions.
- [Custom Endpoints](https://docs.kong.fyi/configuration/custom-endpoints.md): Use local or third-party OpenAI-compatible models with Kong
- [Environment Variables](https://docs.kong.fyi/configuration/environment-variables.md): All environment variables recognized by Kong
- [LLM Providers](https://docs.kong.fyi/configuration/llm-providers.md): Configure Anthropic, OpenAI, or custom LLM providers for Kong
- [Installation](https://docs.kong.fyi/getting-started/installation.md): Install Kong from PyPI or source, configure your environment, and verify everything works.
- [Quickstart](https://docs.kong.fyi/getting-started/quickstart.md): Go from zero to your first binary analysis in five minutes.
- [Setup Wizard](https://docs.kong.fyi/getting-started/setup-wizard.md): Walk through kong setup step by step — providers, API keys, custom endpoints, and what gets saved.
- [What is Kong?](https://docs.kong.fyi/getting-started/what-is-kong.md): AI-powered reverse engineering that recovers function names, types, and structures from stripped binaries in minutes instead of days.
- [kong analyze](https://docs.kong.fyi/reference/cli/analyze.md): Run Kong's full analysis pipeline on a binary
- [kong eval](https://docs.kong.fyi/reference/cli/eval.md): Score analysis output against ground-truth source code
- [kong info](https://docs.kong.fyi/reference/cli/info.md): Display binary metadata without running analysis
- [kong setup](https://docs.kong.fyi/reference/cli/setup.md): Interactive setup wizard for configuring Kong
- [LLM Models & Pricing](https://docs.kong.fyi/reference/llm-models-pricing.md): Complete pricing and model limits reference for all supported LLM providers
- [Supported Architectures](https://docs.kong.fyi/reference/supported-architectures.md): Architecture and language support matrix with confidence levels
- [Analyzing a Binary](https://docs.kong.fyi/usage/analyzing-a-binary.md): Run Kong's full analysis pipeline on a stripped binary
- [Binary Info](https://docs.kong.fyi/usage/binary-info.md): Inspect binary metadata without running a full analysis
- [Evaluating Results](https://docs.kong.fyi/usage/evaluating-results.md): Score Kong's analysis output against ground-truth source code
- [Output Formats](https://docs.kong.fyi/usage/output-formats.md): Understanding Kong's three output formats: source, JSON, and Ghidra writeback

## Optional

- [GitHub](https://github.com/amruth-sn/kong)
- [PyPI](https://pypi.org/project/kong-re/)