Practical overview
When it fits
Use it from Ghidra when you want model-assisted code explanations or a structured way to explore an indexed binary. Configure a compatible API endpoint for LLM-generated summaries and queries.
Who it is for
- Reverse engineers who use Ghidra and want explanations of disassembly or decompiled pseudo-C.
- Analysts who want contextual model queries, binary relationship browsing or searches across generated summaries.
Documented uses
- Ask for explanations of functions and instructions in disassembly or decompiled pseudo-C.
- Ask contextual questions about a selected function or location and continue them in multi-turn chat.
- Index binary structure, browse function relationships and search summaries or security annotations.
- Scan for security-relevant API use and patterns such as network, file-I/O, cryptographic APIs and selected strings.
Requirements & limits
The public repository lists an MIT license. The README describes a Ghidra extension installation and API endpoint configuration.
- Requires Ghidra and an OpenAI v1-compatible API for model-backed features.
- The README names local interfaces including Ollama, LM Studio and Open-WebUI, and cloud providers including OpenAI, Anthropic and Azure.
- Semantic graph indexing and LLM summary generation are separate stages; summary generation requires an API.
- Reasoning controls depend on the model and provider, and higher effort may increase latency and token use.
- The repository is implemented in Java and lists the MIT license.
Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.
Install and connect GhidrAssist in Ghidra
Enable the extension in Ghidra and configure its model endpoint.
Before you start
- Ghidra and a binary to inspect.
- A binary release ZIP if installing from a release archive.
- An OpenAI v1-compatible API endpoint and its API key.
Steps
If needed, copy the release ZIP into Ghidra_Install/Extensions/Ghidra.
In Ghidra, choose File → Install Extension and enable GhidrAssist.
Load a binary, open CodeBrowser, enable GhidrAssist under File → Configure → Miscellaneous, then open it from Window → GhidraAssistPlugin.
Set the API host and key, and check that the RLHF and RAG database paths suit your environment.
Expected outcome
GhidrAssist is enabled in the CodeBrowser and configured to connect to the selected compatible endpoint.
Guide limitations
- Provider setup details differ by endpoint.
- This setup procedure does not establish that a particular Ghidra, model or endpoint combination has been tested.
Index and browse a binary in the Semantic Graph
Create a structural index and explore its function relationships and summaries.
Before you start
- GhidrAssist installed and a binary loaded in Ghidra.
- An OpenAI v1-compatible API configured to generate semantic summaries.
Steps
Open the Semantic Graph tab and select ReIndex Binary to extract structural relationships.
Select Semantic Analysis to generate LLM summaries.
Browse indexed functions in the list, inspect relationships in the graph view, or search summaries and security annotations.
Expected outcome
The tab presents indexed binary structure for list, graph and text-search exploration; generated semantic summaries are available when the analysis step completes.
Guide limitations
- Generating semantic summaries requires an API; graph queries can use precomputed summaries without an LLM call for every query.
- Security feature scans report detected features and assigned risk levels; those results are analysis aids, not a guarantee of vulnerability detection.
Sources & review
Source record details
- Source availability
- Active
- Sources checked
- 7 November 2026
- Original website title
- symgraph/GhidrAssist · Official homepage title · retained in its source language