AgentMonitor
See what your AI agent actually did.
Live at https://cogniguardai.com/ — v0.1.1 is on PyPI: pip install cogniguardai.
AgentMonitor is a flight recorder for AI agents. It records every prompt, tool call, file touched and dollar spent — so when your agent does something weird, you can rewind and see exactly what happened.
- Local-first. Runs as a desktop app on your laptop. Your prompts never leave your machine.
- Free. No signup, no telemetry, no cloud account. Forever.
- Works with what you use. Cursor, Claude Code, OpenAI, Anthropic, LangChain, AutoGen, Smolagents, Ollama — and any custom agent.
Made by CogniGuard AI.
Install
Windows (no Python required) — download
AgentMonitor.exe
(20 MB) from the latest release
and double-click. First launch may show a SmartScreen warning
(not yet code-signed); click More info → Run anyway.
Python (any OS) —
pip install cogniguardai
agentmonitor # opens the dashboard at http://127.0.0.1:8765
The slim baseline is ~174 KB on the wire and pulls only fastapi, uvicorn, pydantic, httpx, rich, PyYAML and pywebview — no torch, no transformers, no Playwright. Optional features ship as extras:
pip install 'cogniguardai[ml]' # interp probes, NLA decoders, Llama Guard
pip install 'cogniguardai[browser]' # Playwright session controller
pip install 'cogniguardai[all]' # everything
Run from source
git clone https://github.com/cogniguardai/agentmonitor.git
cd agentmonitor
pip install -r requirements.txt
python -m agent_monitor.run_server
# then open http://127.0.0.1:8765
What it does
- Record. Every agent run is captured automatically — prompts, tool calls, files touched, tokens spent, money burned.
- Replay. Open any past run and scroll through it like a video.
- Rewind. When something goes wrong, you have the receipts.
Documentation
- Marketing site & live demo: https://cogniguardai.com/
- Issues / feature requests: GitHub Issues
- Email: hello@cogniguardai.com
Repository layout
agent_monitor/ Python package (the recorder + dashboard backend)
marketing/ Source for the marketing site at cogniguardai.com
requirements.txt Python dependencies
License
MIT — see LICENSE.
Privacy
AgentMonitor does not phone home. Ever. All data lives in a local SQLite database on your machine. There is no analytics, no telemetry, no signup, no account.
Metadata
Release files for cogniguardai 0.1.1
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | |
|---|---|---|---|
| cogniguardai-0.1.1.tar.gz | 159.7 kB | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| cogniguardai-0.1.1-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 337.2 kB
Release files / cogniguardai-0.1.1.tar.gz
| Download URL | cogniguardai-0.1.1.tar.gz |
|---|---|
| Size | 159.7 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
2dcf6850574568bd345b411243c2ae46042143697724ce5217dc493823988c90
|
|
BLAKE2b-256 checksum How to use checksums |
fecd52b8d437d53e81d3822bb711a506cf16c7adc4c140650c244115e69f2acc
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.11.9
|
Release files / cogniguardai-0.1.1-py3-none-any.whl
| Download URL | cogniguardai-0.1.1-py3-none-any.whl |
|---|---|
| Size | 177.5 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
2eb18d81ac137cf4d6dadb021798b7a3fa276abe9d97493374d9e7c9a95aae64
|
|
BLAKE2b-256 checksum How to use checksums |
338349873cd77f35a738add2e3a4c5ee9744847dc28cbe0c0801197a534728d1
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.11.9
|