EventWhisper
Enables fast, scriptable access to Windows .evtx event logs for incident response and digital forensics. Supports filtering events by time windows, Event IDs, and keywords with field projection to reduce output size.
README
EventWhisper — A Windows Event Log MCP
EventWhisper exposes fast, scriptable access to Windows .evtx logs via an MCP (Model Context Protocol) server.
It’s pure Python (uses the evtx library), instead of a PowerShell wrapper and does not execute commands on the host — safer for incident response, digital forensics, and threat hunting.
The server exposes two primary tools:
- List EVTX files (optionally recursive) for any directory.
- Filter events to directly search for specific records.
Built specifically for IR/DFIR & hunting so you don’t need to look up Event IDs all day. Designed, developed, and tested on Windows with Claude — and usable from any MCP client.
Features
- Targeted filtering: time window, EventID(s), and case-insensitive keywords (include/exclude)
- Field projection: return only the dotted paths you need (huge output reduction)
- Input normalization: accepts flexible input formats so it still works if types/format aren’t perfect.
- MCP-ready: plugs into Claude Desktop (and other MCP clients)
Demo
https://github.com/user-attachments/assets/e1878c86-56f6-466b-bbc4-04d1edeeb36b
Requirements (Windows)
EventWhisper can be set up with Poetry, whether you run it inside Claude or standalone.
On Windows, make sure Poetry is on your PATH.
Install Poetry (PowerShell):
(Invoke-WebRequest -Uri https://install.python-poetry.org -UseBasicParsing).Content | py -
Add Poetry to PATH (user install):
C:\Users\<YourUser>\AppData\Roaming\Python\Scripts
Verify:
poetry --version
Install & Run
Primarily developed for use with Claude, but you can run the MCP server standalone and plug it into other LLMs.
Clone and install:
git clone https://github.com/hexastrike/eventwhisper
cd eventwhisper
poetry install
Run the MCP server:
poetry run python -m eventwhisper.mcp.server
# (or) plain Python if deps are installed globally:
python -m eventwhisper.mcp.server
Claude MCP configuration
Add this to Claude Desktop’s MCP config (e.g., %AppData%\Claude\claude_desktop_config.json):
{
"mcpServers": {
"EventWhisper": {
"type": "stdio",
"command": "poetry",
"args": [
"-C",
"C:\\Path\\To\\eventwhisper",
"run",
"python",
"-m",
"eventwhisper.mcp.server"
],
"env": { "PYTHONIOENCODING": "utf-8" }
}
}
}
Open Settings → Developer to confirm EventWhisper is registered and “running”.

When prompted, allow the tool:

Try a first prompt:
Use EventWhisper to list
.evtxfiles inC:\Windows\System32\winevt\Logs.
Development
EventWhisper was written to be simple and maintainable. We use pytest for tests and ruff for linting/formatting. Run everything locally (or via pre-commit):
poetry install
poetry run ruff format .
poetry run ruff check . --fix
poetry run pytest --cov=eventwhisper --cov-report=term-missing
# pre-commit on all files:
poetry run pre-commit run --all-files
# pre-push hooks (full tests + coverage):
poetry run pre-commit run --hook-stage push --all-files
When making pull requests, ensure all tests pass and coverage thresholds are met.
Every new utility or feature should include clear tests and follow PEP 8. Tests live in the tests/ directory.
Core layout & API
Primary logic lives in eventwhisper/evtxio/evtxio.py — EVTX iteration/filtering/projection.
The core server function is get_events_from_evtx() (a wrapper around the iterator iter_events_from_evtx()):
def get_events_from_evtx(
provider: str | Path,
start: datetime | str | None = None,
end: datetime | str | None = None,
results_limit: int | str | None = RESULTS_LIMIT,
event_ids: int | str | Sequence[int] | Sequence[str] | None = None,
contains: str | Sequence[str] | None = None,
not_contains: str | Sequence[str] | None = None,
fields: str | Sequence[str] | None = None,
) -> list[str]:
Constants live in eventwhisper/utils/config.py and include both:
RESULTS_LIMIT— max number of events returned (LLMs have token/char caps).SCAN_LIMIT— max events scanned to keep responses fast on huge logs.
Example Python usage
# example.py
from __future__ import annotations
from datetime import datetime, timezone
from eventwhisper.evtxio.evtxio import get_events_from_evtx
# 1) First matching event from Security log
print(get_events_from_evtx(r"C:\Windows\System32\winevt\Logs\Security.evtx", results_limit=1))
# 2) Filter by keywords and project fields
events = get_events_from_evtx(
r"C:\Windows\System32\winevt\Logs\Security.evtx",
contains=["powershell", "Invoke-"],
fields=["Event.System.EventID", "Event.EventData.Image", "Event.EventData.CommandLine"],
results_limit=5,
)
for e in events:
print(e)
# 3) Time-bounded query (UTC)
start = datetime(2025, 1, 1, tzinfo=timezone.utc)
end = datetime(2025, 1, 31, 23, 59, 59, tzinfo=timezone.utc)
print(len(get_events_from_evtx(r"C:\Windows\System32\winevt\Logs\Security.evtx", start=start, end=end)))
Run it with poetry run python example.py.
Troubleshooting
- Big logs: results are capped by
RESULTS_LIMITand scanning is bounded to keep responses fast. If your MCP client doesn’t realize results are capped, run a second, narrower query. - “Blocked” files: if you downloaded
.evtxfrom the Internet, right-click → Properties → Unblock. - Paths: in Python, prefer raw strings for Windows paths, e.g.,
r"C:\path\file.evtx". - README video: GitHub doesn’t render
<video>; use a thumbnail → MP4 link (see Demo section).
Roadmap
- Faster scanning for very large
.evtxfiles - Better normalization for malformed/corrupted EVTX
- Convenience utilities (e.g., quick overview: counts, unique EventIDs, time range)
- Optional paging / multi-query strategies for huge datasets
License
Distributed under the GPLv3. See LICENSE.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.