kiji-proxy
Privacy proxy for your OpenAI requests
Runssource: GitHubGoApache-2.0commit c916161b32a5
Go, Apache-2.0 licensed. The project labels itself: ai gateway, data anonymization, data masking, llm security, pii detection, privacy, privacy protection and privacy proxy.
kiji-proxy runs. An Argusic agent installed it in 26 minutes and hit 9 errors and fixed every one without help on a clean machine with no GPU, and the whole session was recorded.
Installed and verified: Go backend builds and serves /health 200 with the quantized ONNX PII model loaded, masks real emails end-to-end via /api/pii/check, Unix desktop Electron app launches and stays running on the virtual display, Python (2 passed) and Go (512 passed) unit suites are green, and the frontend production bundle builds cleanly.
What the agent ended up with on a clean machine, in its own words. How this is measured
Measured by Argusic on a fresh machine every time. Every number links to its evidence. Argusic Score 100 of 100 (the mean of this project's run scores).
At a glance
Subject data from GitHub, linked at the top of this page, refreshed . Test data by Argusic (CC BY 4.0); every number links to a run page with the full log, the recording, and their sha256 hashes.
What was measured
From the run that decided this verdict. Every line is a recorded value, not an estimate.

What went wrong, in the agent's own words
The agent's own words, worst first, never rewritten into advice.
- System Node 18.19.1 cannot run Babel 8/@tailwindcss native bindings (ESM 'with' syntax, missing oxide-linux-x64-gnu)4 minutes
- Electron install.js fails: requires() of ES Module @electron/get@54 minutes
- @tailwindcss/oxide-linux-x64-gnu optional dependency missing after npm install3 minutes
- No Go toolchain on PATH3 minutes
- Model files were Git LFS pointers (737MB model.onnx absent)2 minutes
- Electron aborts: SUID sandbox helper not configured (no root)1 minute
- Install time
- 26 minutes
- Cold machine to finish
- 27 minutes
- Errors hit and fixed
- 9 hit, 9 fixed with no human help
- How the result was proved
- build/kiji-proxy --version printed 'Dataiku's Kiji Privacy Proxy version dev'; curl localhost:8080/health returned 200 {"model_healthy":true,"status":"healthy"}; curl -X POST localhost:8080/api/pii/check with email text returned masked_message + detected_entities EMAIL confidence 0.997; go test ./src/backend/... all ok (512 PASS); uv run pytest tests/ = 2 passed; electron --no-sandbox . forked main+renderer processes that stayed up >30s with backend ready log line
- Model tokens used
- 1,416,197
- Exact commit tested
- c916161b32a5
- Container image
- sha256:cdd920bce783
What was not measured
The other half, so you can see exactly how far this evidence goes. A gap is printed as a gap and never filled in.
- Peak GPU memory
- this ran on a CPU machine, so there was no GPU to sample
- Behaviour over a long run
- each attempt uses the smallest documented settings and a fixed time budget, so this measures whether it runs, not how it holds up under load
The end of the recording
The last lines of the terminal session that decided this verdict, exactly as they were recorded.
ok github.com/dataiku/kiji-proxy/src/backend/pii (cached) ok github.com/dataiku/kiji-proxy/src/backend/pii/detectors (cached) ok github.com/dataiku/kiji-proxy/src/backend/pii/generators (cached) ok github.com/dataiku/kiji-proxy/src/backend/processor (cached) ok github.com/dataiku/kiji-proxy/src/backend/providers (cached) ok github.com/dataiku/kiji-proxy/src/backend/proxy (cached) ok github.com/dataiku/kiji-proxy/src/backend/server (cached) ? github.com/dataiku/kiji-proxy/src/backend/telemetry [no test files] All verification complete. Final report: tokens used 1,416,197 All verification complete. Final report:
Replay the whole session, every command from a clean machine to this point.
How it was tested
One agent, one container, no human help. It installs the project the way its own documentation says to, runs it, and fixes what breaks. Everything below is recorded as it happened: the terminal session, the log and the exact commit. The full method.
Strengths and limits
Measured facts, not opinions. How this is written.
What went well
- Reached a running state on a clean machine, with the session recorded.
- Recovered from all 9 errors without a human stepping in, which says the failures are documented well enough to solve.
- Ran without a GPU, so it does not need one to start.
- Licensed Apache-2.0, as reported by its host.
What did not, or is not known
- Took 26 minutes to install, slower than the median of the 23 comparable projects Argusic has measured.
- Measured once so far. A single attempt shows it can run, not that it runs reliably.
Of the 24 ai-gateway projects Argusic has installed and timed, kiji-proxy was the 22nd fastest to reach a running state, and 18 of 24 reached one at all.
What it is a reasonable choice for
- Trying it on a laptop or a small server: it reached a running state without a GPU.
Also tested, in the same area
Every one of these was installed and run by Argusic on a clean machine. Nothing appears here that was not tested.
Run history
Topics (from GitHub)
ai-gatewaydata-anonymizationdata-maskingllm-securitypii-detectionprivacyprivacy-protectionprivacy-proxy
Embed the badge
Markdown for the project README. It links back here; terms on the terms page.
[](https://argusic.com/subject/kiji-proxy)Questions
- Does kiji-proxy run?
- Yes. kiji-proxy runs. Argusic installed and launched it on a clean machine in 26 minutes, hitting 9 errors on the way, and recorded the session.
- How did Argusic test kiji-proxy?
- On a fresh, disposable machine, with every command recorded and the repository pinned at commit c916161b32a5. One attempt is recorded, and the full method is on the methodology page.
- What did it cost to test kiji-proxy?
- The run that produced this verdict cost $0.53: real compute and model cost, not a list price per million tokens. It is recorded on that run's page.
- How long does kiji-proxy take to install?
- 26 minutes in the run that decided this verdict, from a clean machine to a running state. That is one measured attempt, not an average.
- Does kiji-proxy need a GPU?
- Not to start. Argusic reached a running state on a machine with no GPU.
- What goes wrong when installing kiji-proxy?
- 9 things broke in the recorded run, and 9 were fixed without human help. Each one, and the time it cost, is listed on this page.
- How does kiji-proxy compare with the alternatives?
- Of the 24 ai-gateway projects Argusic has installed and timed, kiji-proxy was the 22nd fastest to reach a running state, and 18 of 24 reached one at all.
- Where is the evidence for kiji-proxy?
- The recorded run is on this page, each linking to its full log and terminal recording, stored with a sha256 fingerprint so it cannot be quietly altered.